South Africa awards a patent to an AI system / Humans + Tech - #92

+ Why artificial intelligence is being used to write adverts + These algorithms look at x-rays and somehow detect your race + Other interesting articles from around the web


We’ve finally crossed that threshold where a patent has been awarded to a machine. There are challenges to the patent process in many countries around the world. So far, most have ruled that machines cannot hold patents. But, it’s brought up some fascinating arguments with possible repercussions to innovation.

If machines can be inventors, could AI soon monopolise technology?

South Africa became the first country to award a patent to a machine. The patent was granted on June 28 to DABUS (Device for Autonomous Bootstrapping of Unified Sentience), an AI system developed by US scientist Stephen Thaler [Dr Amanda-Jane George, The Conversation].

Thaler says DABUS independently designed a fractal-shaped container for improved grip and heat transfer, and an emergency beacon that flashes more noticeably. So he can’t take credit as the actual inventor. 

He has filed patent applications in 17 countries, as permitted by the international Patent Cooperation Treaty. In the US, UK, Germany, Europe and Australia, the patent offices have not approved them. Patent office decisions are pending in 11 other countries.

Interestingly, in Australia, although the Commissioner of Patents ruled that Thaler could not patent inventions by DABUS, Australia’s Federal Court overturned the decision and ruled that an AI system can be named as an inventor.

Patent law language in each country varies. In the USA, patent law specifies that inventors must be human. There is no specification that the applicant has to be a human in other countries such as Australia.

Another point of contention is that patent offices don’t see how a machine can own what it invents. Ownership is critical in the patent process.

Some have concerns about what it will do to innovation and how it will further increase the dominance of big tech companies if AI systems can be patent holders.

There are concerns that accepting machines as inventors could, as Melbourne Law School senior fellow Mark Summerfield has warned, create an avalanche of “automated patent generators” monopolising technology. 

This would further entrench the dominance of tech companies for whom AI is central, such as Google, Apple, Facebook, Amazon and Alibaba. As University College Cork economist Wim Naudé, has written, these platforms have a huge first-mover advantage in AI, “turning them into monopolists and gatekeepers”.

Thaler is challenging the rulings by the US Patent and Trademark Office and the European Patent Office.

Why artificial intelligence is being used to write adverts

Advertisers are now using software by a company called Phrasee to generate their advertising slogans [Michael Dempsey, BBC].

When Dixons Carphone wanted to push shoppers towards its Black Friday sale, the company turned to Artificial Intelligence (AI) software and got the winning line "The time is now".

Saul Lopes, head of customer marketing at Dixons Carphone, thinks it worked because it didn't have the words Black Friday in it.

His human copywriters had produced dozens of potentially successful sentences but they all mentioned Black Friday. It was technology that broke this chain of thought.

Many agencies and advertisers are currently using Phrasee to assist their copywriters with fresh and unique ideas.

These algorithms look at x-rays and somehow detect your race

Researchers have discovered that artificial intelligence algorithms that read x-rays to assist doctors and perhaps identify issues the doctor may have missed can also determine the patient’s race with remarkable accuracy [Tom Simonite, WIRED].

Evidence that algorithms can read race from a person’s medical scans emerged from tests on five types of imagery used in radiology research, including chest and hand x-rays and mammograms. The images included patients who identified as Black, white, and Asian. For each type of scan, the researchers trained algorithms using images labeled with a patient’s self-reported race. Then they challenged the algorithms to predict the race of patients in different, unlabeled images.

Radiologists don’t generally consider a person’s racial identity—which is not a biological category—to be visible on scans that look beneath the skin. Yet the algorithms somehow proved capable of accurately detecting it for all three racial groups, and across different views of the body.

There have already been many incidences of racial discrimination in health care [Humans + Tech - Issue #32]. This finding raises more concerns that AI can magnify inequality in health care. Even more alarming is that the researchers aren’t sure what cues the algorithms used to identify race accurately.  

Other interesting articles from around the web

Dubai is making its own fake rain to beat 122F heat [Bevan Hurley, Independent]

A project led by the University of Reading in England uses drones to shoot electrical charges into the clouds to get the water drops to merge, stick and form raindrops that fall as rain when they are big enough.

The project tries to get the water drops to merge and stick when they receive an electrical pulse, “like dry hair to a comb”.

“When the drops merge and are big enough, they will fall as rain”, Prof Ambaum told the BBC.

Applying electrical shocks to clouds is preferred as it doesn’t require the use of chemicals.

🏨 Hacker hijacks a capsule hotel’s lights, fans, and beds [Andy Greenberg, WIRED]

Kyasupā, a French hacker, demonstrated how he hacked into a hotel’s network and controlled the lights, fans, and bed at the Black Hat hacker conference.

A TRAVEL TIP: When staying in a “capsule hotel,” the Japanese style of budget accommodation that packs guests into tiny, adjoining rooms not much bigger than their bodies, be considerate of your neighbors. Especially if the capsule hotel you're staying in offers digital automation features—and a hacker is staying in the next room over.


“When I saw all of these features, I thought it was pretty cool, because it means that if I can hack them, I could potentially control all the hotel bedrooms, which is super fun,” Kyasupā wrote to WIRED in a text-message interview ahead of his Black Hat talk. “At the end, I found a total of six vulnerabilities, which allowed me to build an exploit to take control of any bedrooms I wanted from my laptop.”

🤯 The hacker who spent a year reclaiming his face from Clearview AI [Isobel Cockerell, Coda]

Matthias Marx, a hacker and researcher studying security systems, pursued Clearview AI for more than a year to stop using his biometric data without his consent. Clearview AI is a facial recognition company highlighted in Issue #13 of Humans + Tech. Coda interviewed Marx. An excerpt from the interview is presented below.

Why is remaining anonymous in the offline world important?

I think it should be important to everyone. Under surveillance, we change our behavior. If I want to attend a protest, but I know it’s easy to be identified, I might decide not to go, even if it was completely legal to do so. Likewise, I might not want to go to the psychologist’s office, if I knew that I was being identified wherever I went. 

Matthias’ difficulty in getting his face removed from Clearview AI’s database inspired him to start a campaign called Reclaim your Face that aims to ban biometric mass surveillance.

You’re part of a campaign called “Reclaim your Face.” What does that mean?

At the moment we don’t really own our faces. There are already lots of biometric experiments out there that use, say, CCTV to process our faces without our consent. We need to do something if we want to claim our faces back, because at the moment, companies just could use our face to identify us. 

Quote of the week

“AI can't hurt if it generates interesting suggestions, but it's like satnav in a car. Great for directions but you don't allow it to drive the car!”

—Rory Sutherland, Vice-Chairman and veteran copywriter at Ogilvy, from the article, “Why artificial intelligence is being used to write adverts” [BBC]

I wish you a brilliant day ahead :)