Contents
IBM technological visionary Oleg Byakhov shared his opinion on ethical trends in IT
The issue of technology ethics has been around for a long time. And if everything is relatively clear in terms of human responsibility for decisions made by algorithms, then what about the financing of technological developments by the military? What about managing people’s behavior with the help of AI for the sake of business development?
Trend Expert
Oleg Byakhov, Director for Government Programs and Business Development of IBM in our country and the CIS. Graduated from the Faculty of Philosophy of the Rostov State University with a degree in Logic. Co-head of the working group for the preparation of the first Information Society Strategy in our country, head of the first Russian IT foresight (2005-2007). Since 2007 he has been working at IBM.
Ethics of AI algorithms
To date, the most discussed issue in the context of ethics of algorithms is biometric data. In July 2020, IBM removed the entire line of facial recognition products from the market. The fact is that in 2016-2018, researchers found that most commercial facial recognition products that were supplied by the largest companies, including IBM, had a recognition error rate for white men much lower than the error rate for black women. The reliability of recognition of white men is above 97%, and black women – from 65 to 80%.
If such algorithms are built into the law enforcement system, then black women may be unreliably suspected of any crime. To remedy this situation, IBM released a publicly available dataset based on publicly licensed photographic images of a million faces in early 2019. At the same time, standard training datasets include only about 2 face images, and more than 70% of them are white men. And when you train a machine on an incorrect set of data, no matter how good your algorithm is, it will still come to incorrect conclusions.
An important aspect of any use of artificial intelligence is that AI complements human intelligence, not replaces it. The final decision must always be made by people.
In the context ethics of using AI we can list several of our cloud services for media and marketers, for example, IBM Personality Insights, IBM Watson Advertising, IBM Watson Media. They are designed to engage the online audience by personalizing the content or product offer, taking into account the mood of the consumer.
The trend of personalization in ten years will lead to the fact that each person will have the opportunity to receive the most customized, specially made product for him. To do this, artificial intelligence needs to be embedded not only in e-commerce recommendation algorithms, but also in the entire production chain. This is necessary so that the product or service is available at the moment when the user of the system has a corresponding request.
Ethics of AI funding
IBM is leading the development of a new generation of AI – neurosymbolic. It will be trained on the principle of teaching children – that is, the algorithm will be based on the mechanism of the child’s cognitive behavior. The program is sponsored by the Defense Advanced Research Projects Agency (DARPA).
This organization has contributed to the development of a huge number of technologies. For example, the Internet and machine translation systems originated as DARPA projects and are now at the heart of the digital world. That is, even if promising technologies are being explored by the US government, they, as in the case of other countries, can have both civilian and military applications. Companies and organizations cannot completely limit the use of their technologies for military purposes – for example, patented developments become public knowledge. Restrictions on the circulation of technology are imposed by governments.
“America has historically used export controls on advanced technologies developed here at home to ensure they are not misused abroad in ways that are contrary to national security interests, foreign policy priorities, or American values.” Christopher Padilla, Vice President of Government and Regulatory Relations at IBM.
In the field of artificial intelligence, IBM has two global collaborative research centers with major universities: the Massachusetts Institute of Technology (MIT) and the University of Notre Dame in the United States. IBM funds university research because it helps find solutions to fundamental problems that are then used in a wide range of technologies. We also organize internships for young scientists from universities, including Russian ones, at IBM Research laboratories.
Ethics of autonomous systems on water and on land
IBM is only tangentially involved in the development of solutions for autonomous driving, such as Olli’s 3D-printed autonomous bus. The company is currently focusing on maritime vessel management tasks. In April 2021, the fully autonomous ship Mayflower, which was recently launched in the British Plymouth, will set off on a journey.
Maritime management is a more complicated story than paved road management. In whose interests will the autonomous ship move in order to cause the least damage in case of a possible accident? The safety of the AI control algorithm in this case lies in correctly timing sufficient time for a successful maneuver, rather than making a critical decision two seconds before imminent tragedy.
Ethics of the labor market: what will happen to professions in ten years
AI excludes routine tasks, but does not necessarily lead to job cuts. The labor market generates a huge number of digital tasks that require the inclusion of AI in the process itself. For example, in the United States, about 70 thousand vacancies in the field of information security are not filled every year, because these personnel are not enough. But professions such as a driver, where there are a huge number of algorithms worked out to date, may become irrelevant soon enough. Such professions are transforming from mass to expert – jobs for drivers on fixed routes are likely to disappear. Even now, in some quarries, you can see exclusively autonomous dump trucks.
But the professions of a doctor, judge, lawyer by 2030 should not disappear. After all, such professionals do not always work according to simple algorithms, which means that their actions cannot be clearly regulated. Add to this the legislation, which is constantly being improved, and this process is accompanied by a lot of debate. AI will help save a person from the routine, keeping in mind and analyzing a large amount of data – after all, the code of laws, for example, of the Russian Federation, is not able to keep in mind any lawyer.
IBM Ethics Standards
Artificial intelligence in the mass consciousness is associated with consumer applications, virtual assistants, and social network analytical systems, because these are the ones that people encounter in the first place. For IBM, AI algorithms are, above all, a valuable tool for optimizing clients’ business.
The company is guided by three main ethical principles in AI training and data use:
In order to comply with these principles, every AI project at IBM passes through the approval of the AI Ethics Board, which includes both technology specialists and lawyers, experts in psychology and ethics.
Subscribe also to the Trends Telegram channel and stay up to date with current trends and forecasts about the future of technology, economics, education and innovation.