Impacts from Computer Vision
Jaden Zhang December 19, 2023
Society
Impacts
For general society and the culture surrounding travel and automobiles, there is a grand shift towards autonomous vehicles (Hill).
Vehicles are able to see all obstacles and obstructions (pedestrians, objects and other vehicles) and process them simultaneously to produce the safest route (AugmentedStartups).
Careers
Each corporation will want to create a model specific to their own needs, so the job market in computer science would increase (Indeed, uTorontoA)!
For example, a new automobile company might want to create their own CV program that can assess road smoothness and roughness for enjoyable travel.
This was my first introduction to AI when I was 10! I love this video and I think it explains well some of the scary aspects of AI (these can be applied to CV in similar ways).
Economy
Impacts
With autonomous vehicles helping society (like mentioned above), the world’s automobile industry will evolve again; like when horses and biological muscles evolved to cars and mechanical muscles (AugmentedStartups).
With AR using CV, the products and services in VR and the entertainment industry grow
This flow of cash can bring opportunities to local economies (Forbes).
Negative:
Computer vision and machine learning has a big downside in possibly reducing the relevance of human labor and jobs. Computers are gaining the ability to process the physical world without human input.
Solutions
The prevention to the problem of labour relies heavily on workers being able to take advantage of their power before they are replaced (protests and workforce protection laws like those of the Hollywood Writers Strike) (Coyle).
There must also be an increase in jobs that are currently out of reach for computer vision (nurses, teachers, social workers).
Environment
Impacts
Computer vision can detect the emission output from factories, power plants, vehicles, livestock and neighbourhoods to then be used as data for emission reduction plans (LinkedIn).
Computer vision can monitor ecosystems and their health with satellite and surveillance imagery (LinkedIn).
Negative:
The push towards artificial intelligence increases the use of energy.
The current state of the world does not offer great means to acquire fast and reliable energy apart from burning fossil fuels.
Solutions
The best way to combat the negative aspect of burning fossil fuels is to push towards green energy in anticipation of the future (LinkedIn).
There may be no way to stop the AI revolution, so the best solution may be to invest in green infrastructure to keep up with high energy demands.
This would be transitioning to more renewable sources of energy and reducing large industries' usages (the typical plans).
An interesting comparison is the large growth of cryptocurrency mining to AI for energy usage. Cryptocurrency uses a large amount of energy and encourages people to act without environmental stewardship. With some places banning crypto-mining, a similar process can be taken for CV if it runs too far!
Human Health
Impacts
Cancer identifying CV programs are used to detect tumours in MRI’s (Mihajlovic).
Computer vision allows subtle patterns in medical imaging to be picked up, increasing the rate of successful early catches for health complications (AugmentedStartups, Javaid).
Negative:
Computer vision gives humans the ability to augment reality at higher qualities. Computers being able to process images allows for high-tech vision devices! People may fall into isolation if too attached to virtual reality, decreasing mental well-being.
Solutions
The major solution to this problem would be summed up in the fact that more awareness about mental health and technology use should be advertised to the public.
Additionally, the problem should be used against itself. It is possible that AR can actually be used to treat isolation by giving humans a chance to connect online when it is much more difficult to connect in-person (like during the Covid pandemic)!
Careers
"Computer science doctors" may arise due to the interconnection between human health and mechanical health.
With the new development in biotechnology, there is a vast field ready to be explored. For example, nurses who can create programs for patients' microbots to administer medicine may be a future job (Forbes).
Ethics
Negative Implications
Using computers to “see” and analyse visual elements is a prime zone for espionage and privacy concerns in a consumers day to day life (LinkedIn).
Big companies are already using computer vision to access data about a user's social media output for targeted advertising.
Computer vision can be used to discriminate against different groups for insurance or other services.
Having your face logged on security cameras using computer vision leads to leaks in privacy of location and identification (LinkedIn).
Solutions
There will need to be cut backs on the permissions of computer systems. Laws must be established, even though they may take away some benefits of CV (LinkedIn).
Opt-in must be mandatory for service providers that have computer vision, as it should be a system where the user is assumed to give no consent, unless explicitly mentioned otherwise.
The general practice of caution and licensable CV models should be controlled when used at a corporate level.
Careers
Jobs in legal action for and against CV and AI may become commonplace (Srivasta).
There also may be a demand for people to properly evaluate and train CV models to prevent discrimination. However, I predict this job will only exist for a short while before CV models begin evaluating themselves. This approach is like Ian Goodfellow's GAN system of training models by referencing each other (MotionMetrics).
Post-Secondary Requirements for Career Options!
For the ethical, law and data analysis portions of computer vision and artificial intelligence, the requirements of education will be based around:
- Law school in conjunction with the understanding of mechanical minds and computer science (Indeed).
- There will need to be an understanding of the rights of a machine and its users.
- Discriminatory biases will be a part of the job, and a post-secondary education in ethnic studies may be useful to understand the different ways datasets can be trained poorly.
- For this segment, you will most likely require a double major in computer science and any of these subjects: law, ethics, philosophy, racial studies... as these career opportunities are so novel that it’s not something that institutions commonly (currently) provide (Indeed). Osgoode York is innovating with this program!