Will AI blur the difference between private and public sectors?

…there are incredibly powerful non-state actors who are also competing furiously to develop this technology. All of the 7 most important technology companies in the world–Google, Apple, Amazon, Facebook, Alibaba, Tencent, Baidu–are making huge investments in AI, from low level frameworks and silicon to consumer products.  It goes without saying that their expertise in machine learning leads any state actor at the moment.

As the applications of machine learning grow, the interactions between these companies and different nation states will grow in complexity. Consider for example road transportation, where we are gradually moving towards on demand, autonomous cars. This will increasingly blur the line between publicly funded mass transportation (e.g. a bus) and private transport (a shared Uber). If this leads to a new natural monopoly in road transportation should it be managed by the state (e.g. the call in London for “Khan’s Cars”) or by a British company, or by a multinational company like Uber?

As Mariana Mazzucato outlined in her fantastic book The Entrepreneurial State, states have historically played a crucial role in underwriting long term, high risk research in science and technology by funding either academic research or the military. These technologies are often then commercialised by private companies. With the rise of visionary and wealthy technology companies like Google we are seeing more high risk long term research being funded by the private sector. DeepMind is a prime example of this. This creates tension when the interests of a private company like Google and a state are not aligned. An example of this is the recent interactions between Google and the Pentagon where over 4000 Google employees protested against Google’s participation in “warfare technologies” and as a result Google decided to not renew its contract with the Pentagon. This is a rapidly evolving topic. Only a week earlier Sergey Brin had said that “he understood the controversy and had discussed the matter extensively with Mr. Page and Mr. Pichai. However, he said he thought that it was better for peace if the world’s militaries were intertwined with international organizations like Google rather than working solely with nationalistic defense contractors”.

Here is more of interest from Ian Hogarth, via…whoever it was that sent it to me!

Comments

Will the printing press blur the difference between private and public sectors?

'It goes without saying that their expertise in machine learning leads any state actor at the moment.'

The NSA does not need the permission of Google, Apple, Amazon, Facebook to use their expertise. Google cancelling a contract is not the same thing as Google receiving this - https://en.wikipedia.org/wiki/National_security_letter

I get annoyed by prior’s anti-GMU rants as much as the next person, but 100% agree with this take.

I too was surprised to read "it goes without saying" in an otherwise informative piece. The NSA can and will take what it wants from any of the 4 American companies listed. BTW why is Microsoft not in the Top 7? MSFT certainly invests heavily in AI, and they are larger, by Market Cap, than Tencent, Baidu, and Alibaba.

Is it really the case that Government funded research has been such a success? Of course we can find anecdotes of when it has helped, but we don’t actually know if it retarded in other cases, perhaps without the US government crowding out funding of space by private individuals we would have has SpaceX much earlier. Concorde was another example where probably resources were spent on white elephant prestige projects, I would bet that all the smart engineers working on that project would have added much more value to society elsewhere.

'Is it really the case that Government funded research has been such a success?'

In the lead up to and during major wars in the 20th, certainly. For example, the production of penicillin was a major government funded research effort - 'At the onset of World War II, Penicillium notatum, the mold made famous by Alexander Fleming in 1928, was well recognized for its ability to inhibit the growth of certain bacteria in laboratory experiments. The pharmaceutical popularly known as penicillin, however, did not exist. Although several American pharmaceutical firms had examined Fleming’s widely distributed mold, none had continued to develop its potential, and it remained a curiosity. American officials only began to take the compound’s potential seriously in the summer of 1941, after a visit by Oxford scientists Howard Florey and Norman Heatley.

----------------------------------------

The advent of penicillin as a clinical and commercial reality depended on a wide range of biological, scientific, and human resources operating in an open network. Arguing that penicillin would play a critical role in the recovery of manpower, US military leaders mobilized these resources instead of prioritizing economic goals. Mold samples, top-secret reports, and scientists began flying all over the country and, in some cases, the world. This approach induced collaboration among the varied scientists involved with the wartime project, providing them access to a previously unfathomable network of scientific exchange.

Extensive coordination by government agencies made this collaboration possible. The Office of Scientific Research and Development initiated US involvement with penicillin and oversaw most of the scientific work prior to 1943. It coordinated a total of 57 research contracts covering the preliminary studies of penicillin, clinical trials by the Committee on Chemotherapies and Other Agents, and extensive research into the drug’s chemical synthesis. A second agency, the War Production Board (WPB), worked with 21 companies, 5 academic groups, and several government agencies, including the US Department of Agriculture (USDA), to establish large-scale production of penicillin by fermentation. These efforts created a unified scientific workforce comprising a range of mycologists, geneticists, clinicians, chemical engineers, pharmacologists, and chemists working across many sectors.

---------------------------------------------------------

This heightened level of exchange required direct policy interventions by the WPB. By 1944, clinical trials had thoroughly proven penicillin’s usefulness in military medicine, and US strategic planning created heightened demand for the drug. With no foreseeable limit to demand and increasing calls for public and foreign distribution, the corporate leaders who composed the Penicillin Producers Industry Advisory Committee formulated a contract by which the WPB determined what aspects of research and development were most vital to the overall project and distributed this information accordingly. Continuing this effort, the WPB obtained exemptions from the Justice Department for companies after concerns arose that pooling technical information might violate antitrust laws. The following months saw a sharp increase in production, which continued until after the close of the war, as 21 factories individually commenced production, fully equipped with deep-tank fermentation and penicillin refinement capabilities. According to one government official’s count, production rates between May and June of 1944 saw “an increase in monthly production over 250 times attained in one year.” https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3673487/

'I would bet that all the smart engineers working on that project would have added much more value to society elsewhere'

You are aware of where jet engines come from, right? Government funded projects - https://en.wikipedia.org/wiki/History_of_the_jet_engine

"Of course we can find anecdotes of when it has helped"

That wasn't an instruction CP.

And that was not merely an anecdote, particularly if you read the whole thing. To develop mass production of penicillin was a major effort beyond the ability of private industry to master.

The strange thing is, a normal American from the 1970s would not need 'anecdotes' to believe that the government that won WWII and landed men on the Moon was only rarely able to master complex, large scale problems in a way that private industry could never hope to match.

It took a generation to convince Americans otherwise.

As Elon Musk makes clear, SpaceX would not exist without substantial government funding.

Second, SpaceX builds on previous work done by government funding. Books detail the design math based on the actual work of successful designers, and engines were available in scrape yards. The first three failures of SpaceX were not due to rocket science, more like software bugs. The U.S. and Russians spend a lot of money figuring out how to build reliable engines, with lots of failures.

A SpaceX took so long because government funding was cut in the 70s and 80s and that meant playing it safe for the one option budgeted.

SpaceX was one of two selected out of six. The other one failed early so a third candidate was picked which flies half as often as SpaceX Dragon. For crew, there were two out of six or so selected. If one were eliminated, NASA would have no problem selecting a third. The funding would mostly be adding the ISS specific hardware and the bigger cost of satisfying NASA of crew safety.

Consider for example road transportation, where we are gradually moving towards on demand, autonomous cars.

Are we? What is the evidence for this? It looks to me like Google et al have too much cash and they do not know what to do with it. So they have invested it in this. What is the evidence that the consumer *wants* autonomous cars? I don't think there is any.

This will increasingly blur the line between publicly funded mass transportation (e.g. a bus) and private transport (a shared Uber). If this leads to a new natural monopoly in road transportation should it be managed by the state (e.g. the call in London for “Khan’s Cars”) or by a British company, or by a multinational company like Uber?

The distinction between a state-run taxi company, a British copy of Uber and Uber suggests to me that there is no blurring of the line between the private and the public. If the distinction was becoming irrelevant it would not matter which of these three ran London's cars, but the author seems to think that it will matter to the voters of London. As I expect it will. So his second line does not follow from the first.

As Mariana Mazzucato outlined in her fantastic book The Entrepreneurial State,

And there's the problem. What makes this book fantastic except its defense of 1970s-style corportatism which is nostalgic for some older writers of a Leftist bent?

states have historically played a crucial role in underwriting long term, high risk research in science and technology by funding either academic research or the military. These technologies are often then commercialised by private companies.

Where we would be without Concorde. The state's success in under-writing research is none-too-good. However again his conclusion does not follow from his premise. Google paying for, say, the next Hubble telescope is not blurring the line between public and private. Giving Elon Musk control of Mars through a version of the East India Company would. Anyone seriously doing that? The state in the West is increasingly incapable of performing even the most basic of tasks.

"What is the evidence that the consumer *wants* autonomous cars? I don't think there is any."

That's why Sergei and Larry have billions in the bank while all you can do is leave crappy comments on the interwebs.

No. The consumer wants a fast and efficient search engine. That is why Brin and Page have billions in the back and all I can do is leave crappy comments on the internet. They are betting that the public *also* wants cars that drive themselves. The evidence for this is what?

There is a rule in car accidents that nothing the government does makes much of a difference. No road rules, no speed limits, nothing. What determines the number of deaths on the road is the number of deaths on the road. It rises until it reaches a limit and then people drive more carefully. This suggests to me that people are more robust about driving than other people seem to think

The demand for Uber in major urban areas is enormous. If you can cut out the labor component of Uber type services there will be a big ROI on that alone.

Also, eventually people who spend 40 minutes driving to work will not be able to compete with those who work while the car drives itself.

I don't understand your argument in the first place, I believe that a substantial amount of people do not like doing the task of driving and would prefer to substitute paid labor or leisure hours. Why construct a dishwasher, laundry machine, or microwave when there was no proof that people didn't like doing dishes, or cooking (many people still like cooking just as they do driving, but at the very least there is a time and place for new devices).

The life of the average American man-on-the-cul-de-sac is so mundane, banal and meaningless that weaving his Corolla through freeway traffic to and from work is the most exciting part of his day. Giving up that opportunity to intimidate his fellows would turn him into even more of a eunuch than he already is. Autonomous cars seem like a good idea but likely fail on several levels.

Sigh, live in a culdesac and drive an 07 corolla to work. I get a glimmer of pleasure when I cut off a BMW 5 series. Wife refuses to get a new car since our corolla “works fine.”

Uber's going under.

Ironic that an economist, on an economics site, can't even perform the most cursory of financial analyses to form a view:

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2933177

NB the above is nothing new

https://www.nakedcapitalism.com/2016/11/can-uber-ever-deliver-part-one-understanding-ubers-bleak-operating-economics.html

Once cabs no longer exist, Uber can charge twice or more cab fare.

General purpose AI?

There are easier and more productive applications: compliance. https://www.complianceweek.com/glossary/compliance-meets-artificial-intelligence#.WyoQVCCxWUk

"This will increasingly blur the line between publicly funded mass transportation (e.g. a bus) and private transport (a shared Uber)."

There is no line blurring here, at least for people that understand the distinction between public and private. Government threatens taxpayers with prison to force them to fund the public bus. Shared Ubers are funded voluntarily by riders. If the public bus could be similarly self-supported by riders, then there would be no reason for government to run it. If there is any blurring, it is caused by government intruding into provision of private goods, i.e., goods that are either rivalrous or excludable or both. AI has nothing to do with it.

Admittedly, the Google "warfare technologies" case is an example of blurred lines between politics and commerce but, once again, AI has nothing to do with it. We have seen the same blurring in non-AI businesses, where car rental companies, airlines, and banks have all decided to wade into gun politics. Those are all examples of subsets of employees deciding that they can commandeer company resources to advance their own political views even though those companies hold themselves out to be apolitical entities. For example, when Google describes its business to investors, I doubt that it discloses the material fact that it will divert substantial company resources towards studying and advocating defense policy. Google shareholders believe they own a technology firm, not a think tank or political advocacy non-profit. If Google didn't invest substantial resources in studying the "warfare technologies" issue, then that means it made an uninformed decision without understanding all of the national security and privacy aspects. If it did make an informed decision, then it must have diverted resources away from its software business into security policy and advocacy.

Here is Farhad Manjoo's assessment of tech replacing government: https://www.nytimes.com/2018/06/20/technology/tech-companies-conquered-cities.

AI and Big Data will challenge the ability of the managerial class.

If an algorithm can predict market demand, and manage resources and labor as well as an entrepreneur, what value does he bring that a socialized enterprise doesn't? The algorithm doesn't care whether it works for a private investor or state fund.

Actual conversation:

"OK Google"

"Hi, how can I help?"

"Should we separate infant children from their mothers for misdemeanor illegal entry at the US border?"

"..."

(It's been bubbling there for a couple minutes, I think I broke it. And no, this ain't off topic. It's meta on what's important and if AI is anywhere close to helping us with that.)

We should be careful with developing AI, not to sound sci-fi apocalyptic but the scenarios are ever more possible. We should have AI innovation by the private sector but regulated by the State since artificial intelligence will have so much power and control over day to day activities like mass transportation.

Comments for this post are closed