Connect with us

Life

Revolutionising Crime-Solving: AI Detectives on the Beat

Explore the potential of AI in law enforcement with Soze, a system that can analyse vast amounts of data quickly. Discover the benefits and challenges of this technology.

Published

on

AI in law enforcement

TL;DR:

  • AI-powered system, Soze, is being tested by UK police to solve cold cases by analysing vast amounts of data quickly.
  • The system scanned evidence from 27 complex cases in 30 hours, equivalent to 81 years of human work.
  • Concerns remain about the accuracy and potential biases of AI in law enforcement.

In the fast-paced world of technology, artificial intelligence (AI) is making waves in various sectors, including law enforcement. A police department in the United Kingdom is currently testing an AI-powered system that could revolutionise crime-solving, particularly for cold cases. This innovative approach is not without its controversies, however. Let’s delve into the details of this cutting-edge technology and its implications.

The Power of AI in Crime-Solving

The Avon and Somerset Police Department is at the forefront of this technological advancement. They are testing an AI system called Soze, developed in Australia, which has the potential to condense decades of detective work into mere hours. According to Sky News, the AI was able to scan and analyse evidence from 27 complex cases in about 30 hours. This is equivalent to 81 years of human work, highlighting the system’s incredible efficiency.

Gavin Stephens, the chairman of the UK’s National Police Chiefs’ Council, expressed his optimism about the technology. He noted that Soze could be particularly useful for cold cases with vast amounts of material. The system can ingest and assess this data quickly, providing a fresh perspective that could lead to breakthroughs.

How Soze Works

Soze is designed to scan and analyse various types of evidence, including emails, social media accounts, videos, financial statements, and other documents. By processing this data at an unprecedented speed, the AI can help detectives uncover patterns and connections that might have been missed otherwise. This capability makes Soze a valuable tool for law enforcement agencies, especially those facing personnel and budget constraints.

Concerns and Challenges

While the potential benefits of Soze are impressive, there are significant concerns about its accuracy and reliability. AI models are known to produce incorrect results or even hallucinate information. This is particularly problematic in law enforcement, where false positives can have severe consequences.

Advertisement

Stephens also mentioned another AI project that involves creating a database of knives and swords used in crimes. While this could be a useful tool for investigations, it is crucial to ensure that the AI systems are working correctly and without bias.

Bias and Inaccuracies in AI

One of the most concerning aspects of AI in law enforcement is the potential for bias. A model used to predict a suspect’s likelihood of committing future crimes was found to be inaccurate and biased against Black people. This echoes the themes of Philip K. Dick’s “Minority Report,” where predictive policing leads to false arrests and injustices.

Facial recognition technology, another AI application, has also been criticised for its inaccuracies. Minorities have been wrongly accused of crimes due to false positives generated by these systems. These issues are so concerning that the US Commission on Civil Rights has criticised the use of AI in policing.

The Human Factor

It is essential to remember that AI systems are built on data collected by humans, who can be biased and prone to errors. This means that familiar issues are often baked into the AI from the start. There is a common misconception that machines are infallible, but the reality is more complex.

The Future of AI in Law Enforcement

Despite the challenges, the potential of AI in law enforcement is undeniable. Systems like Soze could significantly enhance the efficiency and effectiveness of investigations. However, it is crucial to address the concerns about accuracy and bias before these technologies are widely adopted.

Advertisement

Ensuring Fairness in AI

Before implementing AI systems in law enforcement, it is essential to conduct thorough testing and validation. This includes assessing the system’s accuracy and checking for any biases in its algorithms. Additionally, law enforcement agencies should be transparent about their use of AI and engage with the public to build trust.

Comment and Share:

We would love to hear your thoughts on the use of AI in law enforcement. Do you think systems like Soze could revolutionise crime-solving, or are you concerned about the potential for bias and inaccuracies? Share your experiences and opinions in the comments below. Don’t forget to subscribe for updates on AI and AGI developments.

You may also like:

Author


Discover more from AIinASIA

Subscribe to get the latest posts sent to your email.

Business

AI Just Killed 8 Jobs… But Created 15 New Ones Paying £100k+

AI is eliminating roles — but creating new ones that pay £100k+. Here are 15 fast-growing jobs in AI and how to prepare for them in Asia.

Published

on

AI jobs paying £100k

TL;DR — What You Need to Know:

  • AI is replacing roles in moderation, customer service, writing, and warehousing—but it’s not all doom.
  • In its place, AI created jobs paying £100k: prompt engineers, AI ethicists, machine learning leads, and more.
  • The winners? Those who pivot now and get skilled, while others wait it out.

Let’s not sugar-coat it: AI has already taken your job.

Or if it hasn’t yet, it’s circling. Patiently. Quietly.

But here’s the twist: AI isn’t just wiping out roles — it’s creating some of the most lucrative career paths we’ve ever seen. The catch? You’ll need to move faster than the machines do.

The headlines love a doomsday spin — robots stealing jobs, mass layoffs, the end of work. But if you read past the fear, you’ll spot a very different story: one where new six-figure jobs are exploding in demand.

And they’re not just for coders or people with PhDs in quantum linguistics. Many of these jobs value soft skills, writing, ethics, even common sense — just with a new AI twist.

So here’s your clear-eyed guide:

  • 8 jobs that AI is quietly (or not-so-quietly) killing
  • 15 roles growing faster than a ChatGPT thread on Reddit — and paying very, very well.

8 Jobs AI Is Already Eliminating (or Shrinking Fast)

1. Social Media Content Moderators

Remember the armies of humans reviewing TikTok, Instagram, and Facebook posts for nudity or hate speech? Well, they’re disappearing. TikTok now uses AI to catch 80% of violations before humans ever see them. It’s faster, tireless, and cheaper.

Most social platforms are following suit. The remaining humans deal with edge cases or trauma-heavy content no one wants to automate… but the bulk of the work is now machine-led.

Advertisement

2. Customer Service Representatives

You’ve chatted with a bot recently. So has everyone.
Klarna’s AI assistant replaced 700 human agents in one swoop. IKEA has quietly shifted call centre support to fully automated systems. These AI tools handle everything from order tracking to password resets.

The result? Companies save money. Customers get 24/7 responses. And entry-level service jobs vanish.

3. Telemarketers and Call Centre Agents

Outbound sales? It’s been digitised. AI voice systems now make thousands of simultaneous calls, shift tone mid-sentence, and even spot emotional cues. They never need a lunch break — and they’re hard to distinguish from a real person.

Companies now use humans to plan campaigns, but the actual calls? Fully automated. If your job was cold-calling, it’s time to reskill — fast.

4. Data Entry Clerks

Manual input is gone. OCR + AI means documents are scanned, sorted, and uploaded instantly. IBM has paused hiring for 7,800 back-office jobs as automation takes over.

Advertisement

Across insurance, banking, healthcare — companies that once hired data entry clerks by the dozen now need just a few to manage exceptions.

5. Retail Cashiers

Self-checkout kiosks were just the start. Amazon Go stores use computer vision to eliminate the checkout experience altogether — just grab and go.

Walmart and Tesco are rolling out similar models. Even mid-sized retailers are using AI to reduce cashier shifts by 10–25%. Humans now restock and assist — not scan.

6. Warehouse & Fulfilment Staff

Amazon’s warehouses are a case study in automation. Autonomous robots pick, pack, and ship faster than any human.
The result? Fewer injuries, more efficiency… and fewer humans.

Even smaller logistics firms are adopting warehouse AI, as costs drop and robots become “as-a-service”.

Advertisement

7. Translators & Content Writers (Basic-Level)

Generative AI is fast, multilingual, and on-brand. Duolingo replaced much of its content writing team with GPT-driven systems.

Marketing teams now use AI for product descriptions, blogs, and ads. Humans still do strategy — but the daily word count? AI’s job now.

8. Entry-Level Graphic Designers

AI tools like Midjourney, Ideogram, and Adobe Firefly generate visuals from a sentence. Logos, pitch decks, ad banners — all created in seconds. The entry-level designer who used to churn out social graphics? No longer essential.

Top-tier creatives still thrive. But production design? That’s already AI’s turf.

Are you futureproofed—or just hoping you’re not next?

15 AI-Driven Jobs Now Paying £100k+

Now for the exciting bit. While AI clears out repetitive roles, it also opens new high-paying jobs that didn’t exist 3 years ago.

These aren’t sci-fi ideas. These are real jobs being filled today — many in Singapore, Australia, India, and Korea — with salaries to match.

Advertisement

1. Machine Learning Engineer

The architects of AI itself. They build the algorithms powering everything from fraud detection to self-driving cars.
Salary: £85k–£210k
Needed: Python, TensorFlow/PyTorch, strong maths. Highly sought after across finance, healthcare, and Big Tech.

2. Data Scientist

Translates oceans of data into actual insights. Think Netflix recommendations, pricing strategies, or disease forecasting.
Salary: £70k–£160k
Key skills: Python, SQL, R, storytelling. A killer combo of tech + communication.

3. Prompt Engineer

No code needed — just words.
They craft the perfect prompts to steer AI models like ChatGPT toward accurate, helpful results.
Salary: £110k–£200k+
Writers, marketers, and linguists are all pivoting into this role. It’s exploding.

4. AI Product Manager

You don’t build the AI — you make it useful.
This role bridges business needs and tech teams to launch products that solve real problems.
Salary: £120k–£170k
Ideal for ex-consultants, startup leads, or technical PMs with an eye for product-market fit.

5. AI Ethics / Governance Specialist

Someone has to keep the machines honest. These specialists ensure AI is fair, safe, and compliant.
Salary: £100k–£170k
Perfect for lawyers, philosophers, or policy pros who understand AI’s social impact.

Advertisement

6. AI Compliance / Audit Specialist

GDPR. HIPAA. The EU AI Act.
These specialists check that AI systems follow legal rules and ethical standards.
Salary: £90k–£150k
Especially hot in finance, healthcare, and enterprise tech.

7. Data Engineer / MLOps Engineer

Behind every smart model is a ton of infrastructure.
Data Engineers build it. MLOps Engineers keep it running.
Salary: £90k–£140k
You’ll need DevOps, cloud computing, and Python chops.

8. AI Solutions Architect

The big-picture thinker. Designs AI systems that actually work at scale.
Salary: £110k–£160k
In demand in cloud, consulting, and enterprise IT.

9. Computer Vision Engineer

They teach machines to see.
From autonomous cars to medical scans to supermarket cameras — it’s all vision.
Salary: £120k+
Strong Python + OpenCV/TensorFlow is a must.

10. Robotics Engineer (AI + Machines)

Think factory bots, surgical arms, or drone fleets.
You’ll need both hardware knowledge and machine learning skills.
Salary: £100k–£150k+
A rare mix = big pay.

Advertisement

11. Autonomous Vehicle Engineer

Still one of AI’s toughest challenges — and best-paid verticals.
Salary: £120k+
Roles in perception, planning, and safety. Tesla, Waymo, and China’s Didi all hiring like mad.

12. AI Cybersecurity Specialist

Protect AI… with AI.
This job prevents attacks on models and builds AI-powered threat detection.
Salary: £120k+
Perfect for seasoned security pros looking to specialise.

13. Human–AI Interaction Designer (UX for AI)

Humans don’t trust what they don’t understand.
These designers make AI usable, friendly, and ethical.
Salary: £100k–£135k
Great path for UXers who want to go deep into AI systems.

14. LLM Trainer / Model Fine-tuner

You teach ChatGPT how to behave. Literally.
Using reinforcement learning, you align models with human values.
Salary: £100k–£180k
Ideal for teachers, researchers, or anyone great at structured thinking.

15. AI Consultant / Solutions Specialist

Advises companies on where and how to use AI.
Part analyst, part strategist, part translator.
Salary: £120k+
Management consultants and ex-founders thrive here.

Advertisement

The Bottom Line: You Don’t Need to Fear AI. You Need to Work With It.

If AI is your competition, you’re already behind. But if it’s your co-pilot, you’re ahead of 90% of the workforce.

This isn’t just about learning to code. It’s about learning to think differently.
To communicate with machines.
To spot where humans still matter — and amplify that with tech.

Because while AI might be killing off 8 jobs…

It’s creating 15 new ones that pay double — and need smart, curious, adaptable people.

So—

Advertisement

Will you let AI automate you… or will you get paid to run it?


You may also like:

AI Upskilling: Can Automation Boost Your Salary?

How Will AI Skills Impact Your Career and Salary in 2025?

Will AI Kill Your Marketing Job by 2030?

Advertisement

Author


Discover more from AIinASIA

Subscribe to get the latest posts sent to your email.

Continue Reading

Life

“Sounds Impressive… But for Whom?” Why AI’s Overconfident Medical Summaries Could Be Dangerous

New research shows AI chatbots often turn cautious medical findings into overconfident generalisations. Discover what that means for healthcare communication.

Published

on

AI-generated medical summaries

TL;DR — What You Need to Know

  • Medical research thrives on precision — but humans and AIs both love to overgeneralise with AI-generated medical summaries.
  • New research shows large language models routinely turn cautious medical claims into sweeping, misleading statements.
  • Even the best models aren’t immune — and the problem could quietly distort how science is understood and applied.

Why AI-Generated Medical Summaries Could Be Misleading

In medicine, the golden rule is: never say more than your data justifies.

Clinicians and researchers live by this. Journal reviewers demand it. Medical writing, as a result, is often painstakingly specific — sometimes to the point of impenetrability. Take this gem of a conclusion from a real-world trial:

“In a randomised trial of 498 European patients with relapsed or refractory multiple myeloma, the treatment increased median progression-free survival by 4.6 months, with grade three to four adverse events in 60 per cent of patients and modest improvements in quality-of-life scores, though the findings may not generalise to older or less fit populations.”

Meticulous? Yes. Memorable? Not quite.

So, what happens when that careful wording gets trimmed down — for a press release, an infographic, or (increasingly) an AI-generated summary?

It becomes something like:

“The treatment improves survival and quality of life.”

Technically? Not a lie. But practically? That’s a stretch.

From nuance to nonsense: how ‘generics’ mislead

Statements like “the treatment is effective” are what philosophers call generics — sweeping claims without numbers, context, or qualifiers. They’re dangerously seductive in medical research because they sound clear, authoritative, and easy to act on.

Advertisement

But they gloss over crucial questions: For whom? How many? Compared to what? And they’re everywhere.

In a review of over 500 top journal articles, more than half included generalisations that went beyond the data — often with no justification. And over 80% of those were, yep, generics.

This isn’t just sloppiness. It’s human nature. We like tidy stories. We like certainty. But when we simplify science to make it snappy, we risk getting it wrong — and getting it dangerously wrong in fields like medicine.

Enter AI. And it’s making the problem worse.

Our latest research put 10 of the world’s most popular large language models (LLMs) to the test — including ChatGPT, Claude, LLaMA and DeepSeek. We asked them to summarise thousands of real medical abstracts.

Even when prompted for accuracy, most models:

Advertisement
  • Dropped qualifiers
  • Flattened nuance
  • Turned cautious claims into confident-sounding generics

In short: they said more than the data allowed.

In some cases, 73% of summaries included over-generalisations. And when compared to human-written summaries, the bots were five times more likely to overstate findings.

Worryingly, newer models — including the much-hyped GPT-4o — were more likely to generalise than earlier ones.

Why is this happening?

Partly, it’s in the training data. If scientific papers, press releases and past summaries already overgeneralise, the AI inherits that tendency. And through reinforcement learning — where human approval influences model behaviour — AIs learn to prioritise sounding confident over being correct. After all, users often reward answers that feel clear and decisive.

The stakes? Huge.

Medical professionals, students and researchers are turning to LLMs in droves. In a recent survey of 5,000 researchers:

  • Nearly half already use AI to summarise scientific work.
  • 58% believe AI outperforms humans in this task.

That confidence might be misplaced. If AI tools continue to repackage nuanced science into generic soundbites, we risk spreading misunderstandings at scale — especially dangerous in healthcare.

What needs to change?

For humans:

Advertisement
  • Editorial guidelines need to explicitly discourage generics without justification.
  • Researchers using AI summaries should double-check outputs, especially in critical fields like medicine.

For AI developers:

  • Models should be fine-tuned to favour caution over confidence.
  • Built-in prompts should steer summaries away from overgeneralisation.

For everyone:

  • Tools that benchmark overgeneralisation — like the methodology in our study — should become part of AI model evaluation before deployment in high-stakes domains.

Because here’s the bottom line: in medicine, precision saves lives. Misleading simplicity doesn’t.

So… next time your chatbot says “The drug is effective,” will you ask: for whom, exactly?

You may also like:

Author


Discover more from AIinASIA

Subscribe to get the latest posts sent to your email.

Continue Reading

Life

Whose English Is Your AI Speaking?

AI tools default to mainstream American English, excluding global voices. Why it matters and what inclusive language design could look like.

Published

on

English bias in AI

TL;DR — What You Need To Know

  • Most AI tools are trained on mainstream American English, ignoring global Englishes like Singlish or Indian English
  • This leads to bias, miscommunication, and exclusion in real-world applications
  • To fix it, we need AI that recognises linguistic diversity—not corrects it.

English Bias In AI

Here’s a fun fact that’s not so fun when you think about it: 90% of generative AI training data is in English. But not just any English. Not Nigerian English. Not Indian English. Not the English you’d hear in Singapore’s hawker centres or on the streets of Liverpool. Nope. It’s mostly good ol’ mainstream American English.

That’s the voice most AI systems have learned to mimic, model, and prioritise. Not because it’s better. But because that’s what’s been fed into the system.

So what happens when you build global technology on a single, dominant dialect?

A Monolingual Machine in a Multilingual World

Let’s be clear: English isn’t one language. It’s many. About 1.5 billion people speak it, and almost all of them do so with their own twist. Grammar, vocabulary, intonation, slang—it all varies.

But when your AI tools—from autocorrect to resume scanners—are only trained on one flavour of English (mostly US-centric, polished, white-collar English), a lot of other voices start to disappear. And not quietly.

Speakers of regional or “non-standard” English often find their words flagged as incorrect, their accents ignored, or their syntax marked as a mistake. And that’s not just inconvenient—it’s exclusionary.

Why Mainstream American English Took Over

This dominance didn’t happen by chance. It’s historical, economic, and deeply structural.

Advertisement

The internet was largely developed in the US. Big Tech? Still mostly based there. The datasets used to train AI? Scraped from web content dominated by American media, forums, and publishing.

So, whether you’re chatting with a voice assistant or asking ChatGPT to write your email, what you’re hearing back is often a polished, neutral-sounding, corporate-friendly version of American English. The kind that gets labelled “standard” by systems that were never trained to value anything else.

When AI Gets It Wrong—And Who Pays the Price

Let’s play this out in real life.

  • An AI tutor can’t parse a Nigerian English question? The student loses confidence.
  • A resume written in Indian English gets rejected by an automated scanner? The applicant misses out.
  • Voice transcription software mangles an Australian First Nations story? Cultural heritage gets distorted.

These aren’t small glitches. They’re big failures with real-world consequences. And they’re happening as AI tools are rolled out everywhere—into schools, offices, government services, and creative workspaces.

It’s “Englishes”, Plural

If you’ve grown up being told your English was “wrong,” here’s your reminder: It’s not.

Singlish? Not broken. Just brilliant. Indian English? Full of expressive, efficient, and clever turns of phrase. Aboriginal English? Entirely valid, with its own rules and rich oral traditions.

Language is fluid, social, and fiercely local. And every community that’s been handed English has reshaped it, stretched it, owned it.

But many AI systems still treat these variations as noise. Not worth training on. Not important enough to include in benchmarks. Not profitable to prioritise. So they get left out—and with them, so do their speakers.

Advertisement

Towards Linguistic Justice in AI

Fixing this doesn’t mean rewriting everyone’s grammar. It means rewriting the technology.

We need to stop asking AI to uphold one “correct” form of English, and start asking it to understand the many. That takes:

  • More inclusive training data – built on diverse voices, not just dominant ones
  • Cross-disciplinary collaboration – between linguists, engineers, educators, and community leaders
  • Respect for language rights – including the choice not to digitise certain cultural knowledge
  • A mindset shift – from standardising language to supporting expression

Because the goal isn’t to “correct” the speaker. It’s to make the system smarter, fairer, and more reflective of the world it serves.

Ask Yourself: Whose English Is It Anyway?

Next time your AI assistant “fixes” your sentence or flags your phrasing, take a second to pause. Ask: whose English is this system trying to emulate? And more importantly, whose English is it leaving behind?

Language has always been a site of power—but also of play, resistance, and identity. The way forward for AI isn’t more uniformity. It’s more Englishes, embraced on their own terms.

You may also like:

Advertisement

Author


Discover more from AIinASIA

Subscribe to get the latest posts sent to your email.

Continue Reading

Trending

Discover more from AIinASIA

Subscribe now to keep reading and get access to the full archive.

Continue reading