Top AI stories of 2021

Global investments in AI broke records, AI regulations moved forward, and language models dominated.

Inside.com Part of   Network December 29, 2021

Presented by

Dear Insiders, I hope everyone is having a wonderful holiday season. As we count down to the New Year, I've compiled a list of the top AI stories of 2021. Stay tuned for tomorrow's issue, when we'll give you our predictions for what could transpire in the year ahead. Cheers! Beth

1. Global investments in AI firms surpassed $77.5B this year, up from last year's record of $36B. By the end of Q2, there were already 50 mega-rounds, up 200%+ YoY. Much of the funding went to companies working on autonomous vehicle tech, including a $2.5B funding round for self-driving startup Waymo and $750M and $2B rounds to Cruise Automation and those using AI in drug discovery. The U.S. continues to lead in AI investment, followed by China and the U.K. - VENTUREBEAT 2. Europe released its proposed rules to govern AI systems, which have gone through months of discussions and revisions since April. Adoption of the so-called AI Act has been sluggish as lawmakers pour over the legal and technical implications and try to figure out who's in charge of it all. In October, European Parliament members did give their support for some of the draft rules. This included a total ban on remote facial recognition and other remote biometric surveillance, a ban on "social scoring systems" that purport to rate citizens' trustworthiness, and the prohibition of data-driven predictive policing, when law enforcement claims it can use AI to predict crimes based on data patterns. Meanwhile, civil society organizations are calling for huge revisions, saying the rules don't go far enough to protect against blackbox bias, AI discrimination, and other issues. Discussions will continue over many months, and the parliament still must approve the final iteration, whatever that may be, in a plenary vote. - TECHCRUNCH 3. A U.S. AI commission released its final report to the President and Congress about AI's role in defense and national security, saying that the U.S. is not currently ready to compete with or defend against China in the new era of AI. The National Security Commission on Artificial Intelligence called for billions in spending on AI R&D and stressed the need for the U.S. to invest in AI-enabled weapons and place "wise restraints" on mass surveillance tools such as facial recognition. The commission was chaired by former Google CEO Eric Schmidt, who himself has been heavily involved in AI this year. Schmidt warned that the EU's proposed rules on making algorithms more transparent could harm Europe and said that AI is at a "watershed moment." He also launched a self-funded project, modeled after the Rockefeller Special Studies Project, to strengthen the U.S.'s AI capabilities and global standing. - VENTUREBEAT 4. Alphabet's DeepMind released the most accurate and complete database of the 3D structures of nearly every protein in the human body, which could supercharge drug discovery, new medicines, and more. The AlphaFold Protein Structure Database, which was developed using DeepMind's AlphaFold 2 AI system, can improve our understanding of how proteins work. Some experts have compared its potential impact to the Human Genome Project. "We believe that this work represents the most significant contribution AI has made to advancing the state of scientific knowledge to date," said DeepMind founder and CEO Demis Hassabis. The U.K.-based AI company has open-sourced the model and the code for anyone to find the structure of any protein, with more to be released over time. - NATURE 5. The U.K., Australia, and France all took action to bar facial recognition company Clearview AI from collecting people's data in their respective countries. Clearview's facial recognition tool primarily helps law enforcement match photos of suspects to images in its database. The U.S.-based company is currently appealing a possible $23M fine and finding this year that it violated U.K. data protection laws — CEO Hoan Ton-That has argued that Clearview doesn't actually have an established EU base, meaning it's not subject to the GDPR and regulators have misinterpreted the company's "technology and intentions." Meanwhile, the company still operates in the U.S., and the country's Patent Office just granted approval on a patent for its facial recognition system. - TECHCRUNCH

A message from ROOFSTOCK Diversify your portfolio with Roofstock One. Accredited investors can explore this innovative way to add the single-family rental (SFR) asset class to their portfolios. Choose Tracking Stocks of targeted portfolios with allocations across markets or Common Stock for maximum diversification. For a limited time, invest for as little as $1,000.* Explore an all-new way to invest at roofstock.com/one. * This limited-time $1,000 minimum investment period ends at 11:59:59 pm PT on December 31, 2021, after which the minimum investment will be $5,000. Get Started Today Roofstock One is not a registered advisor, broker, dealer under the U.S. Securities Act of 1933 as amended. Offerings are not registered and are available only to accredited investors. The investments can lose value, are speculative, and are not suitable for everyone. Read offering materials at www.roofstock.com/one and confer with an advisor before investing. Diversification cannot ensure a profit or prevent loss; it’s a strategy to mitigate risks.

6. Google said it's working to alter how it reviews AI ethics papers and "regain trust" after losing AI ethicists Timnit Gebru and Margaret Mitchell, who co-led its Ethical AI team. In February, Google announced a restructuring of its responsible AI efforts, including engineering VP Marian Croak overseeing "a new center of expertise" at Google Research. Croak later announced plans to double the size of its ethical AI team, noting that CEO Sundar Pichai has agreed to raise the operating budget for a team that's working on tackling AI discrimination, harm, and more. Meanwhile, Gebru has launched her own research institute dedicated to the responsible use of AI outside of Big Tech's influence. - VOX 7. The Pentagon canceled JEDI and introduced a new contract, the Joint Warfighting Cloud Capability, which seeks to upgrade its cloud tech and bolster its AI capabilities. The DoD issued solicitations to Google Cloud, AWS, Microsoft, and Oracle to bid and compete for work. It replaces the $10B JEDI initiative that went to Microsoft until the Pentagon canceled it this year amid legal challenges from Amazon. Google's interest is especially significant, given that it abandoned its Pentagon AI initiative, Project Maven, three years ago amid concerns that its technology would be used to support warfare and killings. - FEDSCOOP 8. Zillow made headlines when it chose to shut down its AI-enabled home buying/selling business. In early November, CEO Rich Barton said the company had lost confidence in its AI algorithms to predict home prices and determine how much to pay for them. Zillow had updated its home valuation algorithm regularly since 2006 and claimed a new AI technique would reduce price estimate errors. In Q3 2021, Zillow purchased 9,680 homes using algorithms and has since been trying to offload 7,000 of them. MoxiWorks CEO York Baur may have said it best: "All the AI and machine learning in the world isn't yet up to the task of the complexity of valuing a home in a rapidly changing market." 9. AI tech companies and labs released a number of language-generating models this year, some bigger than OpenAI's popular GPT-3. The models can talk to people, write their own news stories, and autocomplete code, among other tasks. Microsoft updated its PowerApps software to use GPT-3 to turn natural speech into code. In October, Microsoft and Nvidia introduced a natural language generation model, MT-NLG, with 3x more parameters than GPT-3. The model's 530 billion parameters are still less than Google's Switch Transformer NLP model, which has 1.6 trillion parameters, and Beijing Academy of AI's Wu Dao 2.0, which has 1.75 trillion parameters. Like other language models, however, it's still sometimes toxic and biased. This is because it picks up these qualities from the data it was trained on. - MIT TECH REVIEW 10. More Chinese AI companies, including SenseTime and Megvii, were added to the U.S. government's investment blacklist over links to alleged human-rights abuses in China's Xinjiang region. The U.S. says the companies supplied technology to Chinese security agencies involved in the surveillance and detention of Uyghurs and other minorities, who are being held in "re-education centers" that human rights groups say are akin to concentration camps. AI software provider SenseTime says the accusations are unfounded and is moving forward with its $767M Hong Kong IPO despite its inclusion on the list. The other AI firms added this year include facial recognition provider CloudWalk Technology and supercomputer manufacturer Dawning Information Industry. - BLOOMBERG

A message from SEEDINVEST The next era of fundraising online is here SeedInvest is where high-growth startups go to fundraise online. From a Seed round to a $20M+ Series B, founders in SaaS, consumer, robotics, and more are harnessing the power of equity crowdfunding to take their company to the next level. Here’s what 250+ successfully funded companies know about SeedInvest: Equity crowdfunding is complementary to traditional VC - SeedInvest companies have received follow-on funding from Greycroft, Menlo Ventures, Tiger Global, and more. Your fundraise can be a major growth catalyst - We’ve seen measurable lifts in sales, signups, and press mentions from the moment some companies launch their fundraise. You can raise real capital from real people - No longer just for small deals, the average SeedInvest fundraise last year was $3M+. Want to know more or start your application? Start your application

Honorable mentions: Microsoft-backed AI startup OpenAI made its Codex AI coding assistant available in private beta via its API. The machine learning tool can turn written language into computer code for building games and websites. Elon Musk announced that Tesla is developing a humanoid robot, "Optimus," that builds upon its work in AI and self-driving vehicles. The U.S. is among a handful of countries that don't want to fully ban AI-powered lethal autonomous weapons. UN representatives failed to reach a consensus this month on how to regulate the technologies. AI expert Kai Fu-Lee predicted that AI will intensify wealth inequality as it wipes out peoples' jobs, adding that it's already "making a small number of people ultra-rich and making many people jobless." AI systems can't be listed as inventors on patents because they're not sophisticated enough and aren't actual people, a U.S. federal judge ruled in September. Softbank Group began rolling back its robotics ambitions, starting with its Pepper humanoid robots for consumers. Amazon began requiring delivery drivers to sign a "biometric consent" form that allows the company to collect their facial data and more via its AI-powered driver-monitoring system. Two White House science advisers said they want a "bill of rights" to govern AI systems and protect people against faulty, harmful, and biased systems. The Pentagon said it's making AI investments a "top priority," with plans to invest nearly $1.5B in the next five years to speed up AI adoption.

Upcoming Events: January 04 - Venturing in VC with Landon Campbell (Episode #1) (Register Here) January 05 - DAOs and the intersection of tech & media with Kinjal Shah (Blockchain Capital) (Register Here) January 06 - Augmented and Virtual Reality with Timoni West (VP of AR/VR at Unity) (Register Here) January 11 - The Business of Interactive Gambling with Jon Kaplowitz (Penn) (Register Here) January 13 - Education within the Metaverse with Steve Grubbs (VictoryXR) (Register Here) January 25 - Need To Know NoCoder - Leon Young (Cogniss) (Register Here) January 27 - Understanding the Metaverse (Register Here) February 03 - Blockchain Gaming & Play-to-Earn Economies (Register Here) February 10 - The Business of Wellness (Register Here) February 17 - Customer Acquisition in a post iOS 14 world (Register Here) February 23 - Need to Know NoCoder: Derya Sousa (Kianda Technologies) (Register Here) February 24 - Alternative Investments (Register Here) March 03 - The $12B Appetite for Amazon Aggregators (Register Here)

Beth is an Inside Analyst and a former public policy/investigative reporter. She has covered AI, VR, technology, and e-commerce for Inside. When she's not busy writing, she's hiking in Arizona's desert, playing pickleball, or trying out new Mexican food recipes.   Editor Aaron Crutchfield is based in the high desert of California. Over the last two decades, he has spent time writing and editing at various local newspapers and defense contractors in California. When he's not working, he can often be found looking at the latest memes with his kids or working on his 1962 Ford.

With Roofstock One, accredited investors can explore the single-family rental (SFR) asset class.

SeedInvest is the online fundraising platform designed for growth-obsessed startups. Raise from our network of 575,000+ investors.

Click to Share ➞

767 Bryant St. #203, San Francisco, CA 94107 Copyright © 2021 Inside.com

Did someone forward this email to you? Head over to inside.com to get your very own free subscription! You received this email because you subscribed to Inside AI. Click here to unsubscribe from Inside AI list or manage your subscriptions.