Worldwide spending on AI technology is expected to more than double over the next four years, from $50b now to $110b by 2024, according to a new International Data Corp. forecast. Yesterday, the firm released its latest Worldwide Artificial Intelligence Spending Guide, which predicts that retail, health-care, and banking will see the biggest demand and investments in AI moving forward.

More:

• The report says the primary ways AI will be used are through customer service chatbots, sales automation, and automated threat detection and prevention.
• Companies will adopt AI "not just because they can, but because they must,” said IDC's program president of AI, Ritu Jyoti.
• Retailers are expected to spend $19.6b on AI by 2024, an increase from $8.4b in 2020. They will focus on using AI to improve customer experiences via chatbots and recommendation engines.
• Banks are expected to spend $14.5b by 2024, up from $7.1b in 2020. Most of the focus will be on fraud analysis and investigation.
• Health-care providers are expected to spend $7.55b in four years, up from $3.58b on AI in 2020. They'll primarily use the technology to improve patient care and operations.

Semiconductor startup Kneron introduced its latest AI chip for battery-powered devices. The Kneron KL 720 SoC edge AI chip can handle tasks like speech and image recognition.

More:

• The chip processes information and performs tasks like facial recognition directly on a device, such as a computer, camera, or drone, which keeps customer data off the cloud.
• Edge computing allows data to be stored and processed locally, which can help "internet-of-things" devices run faster and more securely.
• Kneron claims the chip is four times as efficient as Google's Coral and twice as energy-efficient as Intel's Movidius AI chips.
• San Diego-based Kneron recently raised $40m in a Series A round. It's backed by venture capital arms of Alibaba Group, Sequoia, and Qualcomm.
• The edge AI industry is expected to be worth $3.24b by 2025. 

Related:

• Last year, Kneron researchers tricked facial-recognition software at an airport and other locations.
• They used a printed mask of another person's face to fool the payment tablets, which were run by the Chinese companies Alipay and WeChat, as well as facial recognition software at the Chinese border checkpoint and a passport-control gate at Amsterdam's Schiphol Airport.
• However, they were unable to fool Apple's Face ID or Huawei's system, which both use structured light imaging, they said. Kneron CEO Albert Liu said companies "should be held accountable" if they don't protect customers.

Podcast Notes: Fighting bias in AI algorithms

Every Thursday, we summarize a recent podcast to save you time and keep you up-to-date. This week features Elham Tabassi, chief of staff of NIST’s information technology laboratory. She speaks with Tom Temin of the Federal Drive podcast about ways to address unwanted bias in AI algorithms as more federal agencies adopt the technology. [Note: Questions and answers were edited for brevity and clarity.]

Tom: Define what AI bias actually means.

There really isn't one good definition for bias yet, Elham explains.

Researchers are working on this and trying to find an answer. Often, we all use the same term meaning different things. Bias is one of those terms, and the International Standards Organization, ISO, has a subcommittee working on standardization of bias.

NIST has also been doing a literature survey trying to figure out how it has been defined by different experts, with a goal of coming up with a shared understanding.

The current sees bias in terms of disparities in error rates and performance for different populations, devices, or environments. If you have different error rates for different subpopulations, such as in face recognition, that’s not a good bias and something that has to be mitigated.

Another example would be car insurance, where younger people pay at a higher insurance rate than people in their 40s or 50s. The difference in error rate is not bias on the intended behavior or performance of the system. It’s something that’s problematic.

Let’s not forget about human biases, which is one source of bias in AI systems. This can creep into algorithms in different ways because AI systems learn to make decisions based on the training data, which can reflect biased human decisions or historical or societal inequalities.

Other times the bias creeps in because the data is not representative of the whole population (in sampling, one group is overrepresented or underrepresented). Another source is the design and modeling of the algorithm...

Turing, a startup that connects tech companies with remote job vacancies with developers worldwide, raised $14m. The seed round was led by Foundation Capital, with individuals participating including the first CTO at Facebook and also the co-founder of Quora, the founder of Upwork, executives from Google, Facebook and Amazon, and more.

More:

• Turing uses a Machine Learning system to vet and match developers, which it partly attributes towards its success. "Recent fast growth has resulted in more performance data, which has in turn led to rapid improvements in Turing’s vetting and matching accuracy," said CTO Vijay Krishnan. "The result has been even faster growth in both the number of developers on the platform and the number of customers."
• Turing raised $17,000 to $10 million in annualized revenue since becoming generally available one year ago.
• The company has around 150,000 developers on its books that now work in senior positions at the likes of Lambda School, VillageMD, Ohi Technologies, Nexxus Events, and others.

A version of this story first appeared in our newsletter Inside Dev.

Did you enjoy today's Inside AI? We'd like to offer a quick reminder about the benefits of our Premium membership, which helps support this newsletter that relies heavily on reader loyalty.

By subscribing, you'll receive Inside AI ad-free six days per week, along with these special features: 

• New emerging AI technologies.
• GPT-3 free websites/programs.
• Seven ways AI improves business processes.
• AI Masterclass: Trust in AI is key.

And more...

For a limited time, you can try it out for free for 14 days. Click here to sign up!

QUICK HITS

• Astronomers used deep learning to detect as many as 560,000 galaxies in telescopic images.
• Peter Thiel's data analytics company Palantir Technologies filed to go public and be listed on the New York Stock Exchange.
• Verta, an AI/ML model management and operations platform, raised $10M in funding.
• Delight your remote team with curated gift boxes by SnackNation. Get $10 off per box until 8/27!*

PS: We are looking for Business Researchers for our Toronto office. Join us!

• Business Researcher: ECommerce
• Business Researcher: Growth & Advertising
• Business Researcher: UX, Design, & Developer

*This is a sponsored post.