Welcome back. Increasingly, the cook in the kitchen of packaged-food companies is artificial intelligence, enlisted to analyze food data down to the molecular level in the hope of creating products that can satisfy discerning palettes. But taste is personal and complex, and humans are expected to remain heavily involved.

Also, some technology firms are using artificial intelligence to better understand the U.S. electorate; and, the incoming administration is expected to prioritize AI along with quantum sciences and a host of emerging technologies.

Whole Foods Market Inc. this month started selling a plant-based milk in its U.S. stores dubbed NotMilk that was designed using artificial intelligence, Vinod Sreeharsha writes for The Wall Street Journal.

How it was made. Developed by Chilean startup NotCo, the company's machine-learning tool taps into databases related to thousands of plants and plant-based ingredients. By analyzing the molecules in the food, it learns which combinations make, say, cow’s milk and then generates formulas to match. NotMilk is made from cabbage and pineapple, which wasn’t something the company expected when it started the process.

McCormick Co., Conagra Brands Inc. and PepsiCo Inc. are among the food giants using AI to cook up new concepts, WSJ's Jaewon Kang reports, including bourbon pork-tenderloin seasoning and pudding flavors meant to call to mind unicorns.

New suggestions. McCormick’s technicians might not stray far from oregano, basil and other Italian herbs in designing a new pizza seasoning, but AI suggested cumin as an addition because it had proved popular in other relatively new seasoning formulas McCormick had created.

Limitations. Taste is personal and complex, and food giants continue to rely heavily on testing and feedback by human staffers when it applies any AI recommendations.

Bonus: The automat returns, with AI. The automat, a restaurant concept in which customers could buy a hot meal offered from a vending machine’s glass locker, is getting a reboot, WSJ’s Emma Tucker reports. One proprietor plans to open a Manhattan automat this month and use robotics and AI to help manage kitchen operations.

Under a Joe Biden presidency, AI and quantum information science are likely to remain priorities for the U.S., WSJ's Sara Castellanos reports. But his administration’s spending and regulatory approaches are expected to differ from that of President Donald Trump.

Spending. The Biden campaign has proposed innovation funding of $300 billion over four years. That’s in addition to federal research and development spending, which was about $134.1 billion in the Trump administration’s 2020 budget.

• The Biden campaign hasn't specified spending plans for AI and quantum information sciences. The Trump administration said this summer it was on track with its pledge to increase AI funding to more than $2 billion and quantum funding to $860 million by 2022.
• A Biden administration is likely to invest more on emerging technologies, in part because overall research-and-development spending is expected to be higher, said Robert D. Atkinson, president of the nonpartisan Information Technology & Innovation Foundation.

Good calls. Unanimous.ai ran a live online survey in September asking people who would win battleground states, using machine-learning algorithms to gauge prediction confidence and aggregate answers. So far the AI-powered survey is 10 for 11 on battleground race predictions.

And bad ones. Advanced Symbolics Inc.'s AI system predicted Mr. Biden would win. However, it had him taking Florida and winning the electoral vote 372 to 166. The AP has called Florida for Mr. Trump and the AP’s electoral vote totals so far are much closer.

A group of Yale University researchers found that an abused robot is more likely to see humans intervene on its behalf if other nearby robots show sadness for the victim robot, Bonnie Miller Rubin writes for The Wall Street Journal. As robots and attacks on them have increased, the researchers designed an experiment to see how robots might influence human reactions.

The findings. Nine out of 15 people strongly intervened in response to the abuse when the bystander robots expressed sadness. Only three out of 15 intervened when the other robots didn’t react.

Application. “The more we understand human/robot interaction, the more we will be able to design these machines so they can work better for people,” said Marynel Vázquez, assistant professor of computer science at Yale and co-author of the study.

Advancing electric-vehicle batteries. AI is playing a role in suggesting new materials and dramatically reducing research-and-development timelines for electric-vehicle batteries. Within five years, experts say, electric vehicles, or EVs, will reach price parity with traditional combustion-engine cars, thanks in part to AI. (WSJ)

Bias remains a challenge. The AI world is making a strong push to root out this bias, including researchers identifying when AI makes unfair judgments and large tech companies introducing tools to mitigate the issue. But the efforts face obstacles, including little agreement about what “fairness” exactly looks like and the fact that, in some cases, increased fairness comes with decreased accuracy. (WSJ)

Algorithmic substitute for bail system opposed. California voters last week rejected a ballot measure that would have replaced the state’s cash bail system with an algorithmic risk-assessment tool to determine who could be released from jail. Supporters of the measure, including Democratic Gov. Gavin Newsom, argue that cash bail props up a system of racial disparity. Opponents, including Human Rights Watch, contend that the inherently biased algorithm would disproportionately affect people of color. (Sacramento Bee)

Facial-recognition tech banned in Portland, Maine. Voters in the city approved a ballot measure to prohibit the use of the technology by law enforcement agencies and public officials. The ban was originally put in place in August by the city council, subject to the referendum, and now can’t be touched for at least five years. (Engadget)

U.K. to lean on AI to help monitor Covid-19 vaccination effects. The U.K.’s top health regulator signed a contract with software company Genpact for a machine-learning system that can monitor high volumes of potential adverse drug-reaction reports from vaccine takers. (FT)

Facebook, Apple, Google and other tech giants are bracing for continued attention on their operations in a Biden administration, with issues such as competition, innovation and social-media platforms’ policing of content likely to be in focus. (WSJ)

General Motors plans to hire 3,000 IT workers, including offering more remote-working opportunities than it ever has in its effort to expand the workforce and tap talent outside of Detroit. (WSJ)

Write to Enterprise Technology Editor Steven Rosenbush and Pro AI Deputy Editor Angus Loten.

Follow us on Twitter: @wsjproai, @steve_rosenbush, @J_J_McCormick, @tomloftus, @SCastellWSJ, @JaredCouncil  and @angusloten.