Businesses would be wise to pay attention to the A.I. plans that many governments worldwide publish. Savvy companies can use the information to decide how to best invest in A.I.-related initiatives.

Recently, the Brookings Institution think tank and members of Queensland University of Technology analyzed the A.I. plans of 34 countries in an effort to explore the variation among the different plans. The analysis builds upon a prior policy paper from the authors that determined which industries governments believe would be most likely to be transformed by the technology, among other findings.

The winner: health care, which appeared in 28 plans.

The Czech Republic called for using A.I. to streamline several administrative tasks relates to healthcare, such as reimbursement processes, the policy paper said. Meanwhile, Portugal pointed towards A.I.’s potential to help spur medical innovations, like quicker drug discovery and more accurate diagnosis of diseases. 

These countries could be important places for healthcare companies developing A.I.-related tools to invest in.

One important takeaway by the Brookings researchers was that several democratic countries that lack strong home-grown A.I. tech development focus much of their plans on issues like ethics and data governance. The problem for these countries, which included India, Mexico, and Spain, is that any such policies are likely to be insufficient to address the realities of A.I. in their countries, which lag other others in regards to suitable technical infrastructure.

As other researchers have noted, while India has many technologists, the county lacks a capable cloud computing infrastructure that experts believe is critical for A.I.s’ development.

“What we are trying to say here is that A.I. is like running and, before a baby can run, he needs to crawl and walk first,” said Kevin Desouza, a Brookings Institution nonresident senior fellow and Queensland University of Technology professor who co-authored the reports. “Those countries that are lagging technically have never learned how to crawl or walk and so they need to learn how to crawl and then walk before trying to run.” 

Another area often mentioned in the government plans is manufacturing, as the authors previously described. For instance, Desouza and his co-authors identified New Zealand and Finland as two countries emphasizing how A.I. could improve manufacturing. Finland’s A.I. plan, for instance, describes how A.I. will lead to “agile production facilities,” presumably referring to more cutting-edge manufacturing centers that can more quickly produce electronic devices and automobiles.

Perhaps executives leading companies developing advanced factory robots should start learning some Finnish.

Jonathan Vanian 
@JonathanVanian
jonathan.vanian@fortune.com

A.I. IN THE NEWS

So many roads. Self-driving car firms Waymo and Cruise have applied for permits from the California Department of Motor Vehicles to let them operate robotaxis in San Francisco, according to documents viewed by Reuters. The companies did not reveal when they plan to debut their autonomous taxi services, and California’s DMV has not yet approved the permits, the report said. The report said that Waymo plans to debut its robotaxi service with “drivered operations,” presumably meaning that human safety drivers will be in the vehicle. Cruise, on the other hand, is “expecting to deploy vehicles without humans behind the wheel,” the report said.

A.I. to tackle Alzheimer’s disease. Exscientia, a startup specializing in healthcare A.I. technology, said that an Alzheimer’s disease drug candidate, which the company used A.I. to help develop, is entering Phase I clinical trials, an important milestone in the drug discovery process that typically involves human participants. The startup collaborated with Japanese pharmaceutical company Dainippon Sumitomo Pharma to develop the drug. The startup now has three A.I.-designed drugs that have reached Phase I clinical trials.

Is Tesla misleading? The California DMV is currently reviewing Tesla to determine whether Elon Musk’s electric vehicle company is misleading customers by promoting its cars as having “full self-driving capability,” The Los Angeles Times reported. The report notes that Tesla’s technology does not meet “widely accepted engineering standards” of what is deemed fully autonomous driving. From the article: “If the DMV finds Tesla is misleading customers, potential penalties include suspension or revocation of DMV autonomous vehicle deployment permits and manufacture and dealer licenses, the DMV spokesperson said.”

The near future looks bright. Bright Machines, a startup focusing on robotics and manufacturing, will go public by merging with a so-called special-purpose acquisition company SCVX Corp. in a deal that will generate $435 million for the company, The Wall Street Journal reported. From the article:The San Francisco-based company currently has about 25 customers, according to its website. Its existing investors include BMW iVentures—the venture-capital arm of auto maker BMW AG—Eclipse Ventures and Lux Capital.

Smile for the camera. The Financial Times published an interactive article in which readers can test facial-recognition software that attempt to deduce people’s emotional states. It’s a fun activity and shows readers that these emotion-detecting technologies often produce inaccurate results. Still, businesses like advertisers and healthcare providers want to use the technology “to build a profile of you from your expressions.”   

EYE ON A.I. TALENT

Technology executive Hugo Barra is stepping down from his position as Facebook vice president of virtual reality and plans to “explore the healthcare technology space,” he wrote in a Facebook post. Barra was previously the vice president of global for Xiaomi and a Google executive involved with the search giant’s mobile and Android businesses.

Neural Magic, a startup specializing in machine learning development tools, picked Brian Stevens to be the company’s CEO. Stevens was previously the vice president and CTO of Google’s cloud computing unit and the executive vice president and CTO of Red Hat.

JPMorgan Chase & Co. hired Vadim Kutsyy to be the financial firm’s head of A.I. platform solution engineering, trade publication efinancialcareers reported. Kutsyy was previously a distinguished architect of data science and machine learning at PayPal.

Several Waymo executives are leaving the company, including chief financial officer Ger Dwyer and head of automotive partnerships and corporate development Adam Frost, TechCrunch reported. The executive departures follow the recent exodus of CEO John Krafcik from the self-driving car company.

EYE ON A.I. RESEARCH

A.I. to learn to code. Researchers from IBM Research and the MIT-IBM Watson AI Lab published a paper and related dataset intended to help researchers create deep learning systems that can create software code. Similar to how neural networks have become capable of producing more realistic computer-generated text, the researchers believe that the software can create functional code. But producing such software code will only be possible if researchers can train the A.I. systems on the appropriate data.

That’s where IBM’s new dataset comes into play. The dataset contains 14 million code samples and about 500 million lines of code in 55 different programming languages. The researchers said in the paper that the dataset can be used to create deep learning systems that can translate different computer languages,

Research like this could pave the way for more efficient methods for companies to test software, analyze code, and potentially build apps.

FORTUNE ON A.I.

How to hire A.I. talent in a red-hot market—By Jonathan Vanian

Good enough for government work? General Catalyst to back startups tackling the U.S.’s biggest problems—By Lucinda Shen

Whose bodies do we see?—By Ellen McGirt  and Jonathan Vanian

What is SafeMoon? Your guide to the cosmic-themed cryptocurrency—By  Danielle Abril

Andreessen Horowitz just invested in this A.I. health care startup—By Jonathan Vanian

BRAIN FOOD

This Old House. The rise of natural language processing technology, in which computers can understand and generate text, has led to several interesting, new startups. One such example is the startup Listing AI, which CNN recently profiled. The startup, which built its technology on the GPT-3 language model created by the A.I. research firm OpenAI, attempts to help people create A.I.-generated real estate descriptions to help sell homes. People give the startup some basic information about their homes, and the company’s technology automatically generates a description. Still, the author notes that “the results still need work.”

From the article:

The real-life Oakland, California, home that fits with the above description (which my family is currently selling) actually has a pressed tin ceiling in the dining room, rather than the living room, for instance. The descriptions Listing AI created for me are not nearly as specific or well-written as the one crafted by our (human) realtor. And I had to provide the website with a lot of information about different rooms and features of the house and the outdoor landscaping — a process that felt a bit like real-estate Mad Libs — before the website was able to come up with several different descriptions.