Artificial intelligence may inherit, or even amplify, the biases of its creators, who are often mostly male. Uma Iyer finds out what the future holds for AI and women in society.
What is the ‘gender’ of Artificial Intelligence? How do we as a society perceive sentient technology?
If we look at our art, particularly science fiction, whether it be Robin Williams in Bicentennial Man or Rajinikanth in Robot, it is evident that AI is almost always seen as ‘male’.
However, on the flip side if one looks at assistance apps for navigation or Interactive Voice Response Systems in a telephonic services, you notice the active use of a female voice – invariably called a Rachel or a Karen or a Susan!
While technology is moving at a rapid pace, and making life easier in the process, it is still afflicted by the same gender-bias that dominates the current information technology diaspora.
Is AI gender-neutral?
While on the face of it, AI appears neutral, it is in fact driven by humans, which means, it internalizes the same bias we see in the outside world!
We can catch a glimpse of this threat through a small and interesting experiment using Duo Lingo and Google Translate on desktop.
We tried some gender-associated phrases in English:
She is the President
He is cooking
He is a nurse
She is driving a truck
She is a leader
If you translate these into Turkish, and then back to English, you get the following results:
He is president
She’s cooking
She is a nurse
He’s using a truck
He is a leader
“In a 2018 study conducted on Google Translate, translation from English to Turkish, the gendered pronouns were lost as Turkish interestingly is a gender-neutral language. “Translating back to English required the AI-based translation algorithm to analyze and associate a gender with the type of role being described. The data that is used
to teach the AI has strong gender stereotypes. The AI therefore decided it must be HE who is president, driving a truck or a leader, and it must be a SHE who is cooking and is a nurse. AI algorithms are software that is written to learn and create associations in large amounts of existing data. The algorithms’ decisions are as good as the data they utilize for training and understanding. If an algorithm is given a larger amount of text or pictures depicting women doing chores at home and men doing work outside the home, it will increase the statistical probability of these associations and will learn to use them for its future decision-making processing.”
-Dr Muneera Bano (Source: www.kcl.ac.uk)
Is AI suppressing the voice of women?
A real world built and designed using data for men ignores the needs of half its population. This holds true even when artificial intelligence is harnessed to solve challenges facing all of humanity. The default human at the center of most data is ‘Reference Man’ said Caroline Criado-Perez, author of the book Invisible Women: Exposing Data Bias in a World Designed for Men. This ‘Reference Man’ is Caucasian, 25-30 years old, and weighs 70 kilograms, and has been the basis for research studies across sectors for decades. The gender data gap is this “phenomenon whereby the vast majority of information that we have collected globally and continue to collect – everything from economic data to urban planning data to medical data – have been collected on men,” Criado-Perez said during her Breakthrough Days keynote as part of the ‘AI for Good Global Summit 2020’.
Is AI out to get us?
It is believed by many that a sentient synthetic intelligence would be dangerous for humanity. Many thinkers, including the late Stephen Hawking who are concerned that AI could threaten humanity, could cite Facebook’s AI algorithms developing their own language as a test case for the risk posed by such an independent super-intelligent machine.
However, these are more cultural fears and a representation of machine learning in general filtered through our personal dread. Most chatbots are designed to carry out a limited set of functions, so they can be quite limited and boring. The villains of fiction do not necessarily translate into real life tyranny.
Facebook did decide to limit the rollout of its messenger chatbot platform after it found that many of the bots where unable to understand 70 per cent of its user queries. While they are programmed to be humanlike, it is difficult for them to make complex decisions required to plot a rebellion like in the 1984 classic The Terminator.
Enterprises face several ethical challenges in their use of AI technology including Explain ability, Responsibility, Fairness and Misuse.
What’s in the future?
AI is here to stay. More and more people are looking for options to build the ability to understand the human language and interpret and communicate in automated ways for transactional activities. AI chatbots give a great opportunity to work for humans to take away mundane activities. Amazon echo, Siri, Google Assistance, Google home, Amazon Alexa have advanced at an incredible rate to replace the need for reminders or anything to do with touch devices.
The use of Artificial Intelligence could offer as much as $340 billion in cost-saving opportunities for retail companies able to scale and expand the scope of their existing deployments, according to a new global study. The news comes as retailers in the UK and across the world battle against falling sales figures and a crunch in costs related to running stores, while ecommerce continues to eat into the market share of bricks-and-mortar entities.
Voice AI is a conversational tool however that uses voice commands to receive and interpret directives.
Right from children using Alexa seamlessly, voice AI ability interpret basic conversation and are able to rationally help people in need. Retirement homes and especially abled people with autism who take solace in binary and repetitive conversations. A new study shows that implementing an artificial intelligence based medical device in primary care settings could potentially help diagnose autism spectrum in people.
It empowers people with physical disabilities allowing them to take charge of their lives and helps people with neuro-diverse conditions. Generally, people in the neuro-diverse spectrum tend to like consistency and like to take things literally and do tend to like routines. AI does tend to reciprocate these requirements and are built to provide consistency, removing the erratic emotions out of the way.
Machine Learning is always learning! If we check Google Translate on mobile device, it now shows both male and female variants in the text searches conducted earlier, suggesting that machine learning is still learning, and is always open to feedback! This is a great space to be where the better the data input that is fed, we get a holistic output! However, now we have to wonder how a non-binary would like to be viewed in this translation, so there is always scope for improvement and inclusion!
AI continues to find its footing in a variety of business environments. The technology basically gives AI the ability to train a machine to learn. Earlier, we were working toward doing transactional tasks with machines, now we are helping the machines find its voice without having to program. Developers and business alike are finding new avenues for Machine Learning and it’s a field that we must continue to watch out and adapt to as it is here to stay!
In case you missed:
- What is AI going to cost us?
- ‘Art’ificial Intelligence
- Virtual Assistants: A little big thing
- Identifying Human vs. AI Output, Hacks to Beat the Test
- An interview with Google Labs
- Chatbots – The Evolution
- Geriatric Technology: How science makes old age comfortable
- AI and finding love online
- Assisted Intelligence – Are Assistants a thing of the past?
- What Are GPT Stores Really?
9 Comments
i looked at the article at a high level
i think most people who really know AI will disagree with it
AI is gender neutral
you can cherry pick any example of gender specific AI
i can give you female centric AI-based movies.. you talked about Rajnikanth in Robot.. but part 2 features Amy Jackson also as a robot
there’s Ex Machina, Terminator 3 etc. to name a few that are about females as robots/cyborgs/Androids.. not to mention the Borg race in Star Trek.. along with actual conversational physical female robots
you can pretty much cherry pick anything to make a point which is not essentially true
the Twitter stats are all based on real data.. AI only captures real data and extrapolates based on captured data sets.. it does not know gender
it is true that Males as opposed to Females are more interested in getting into the field of AI based development.. and that’s purely based on their interests.. and the outcome is not biased in any way..
although some Vision algorithms are racist
they recognize white hands better than black ones..
but that’s due to the extra light required to detect black hands by the camera.. but this has also been acknowledged and is being worked upon
voice assistants come with an option for a male voice as well as a female voice, based on user’s preference.. and the results produced are the same regardless
this article looks like something a feminist activist would write
Hi Captain VR,
Great insight. The fact that the world is taking an active interest in all aspects to have a feminine point of view comes from all the noise feminists (male/female) have made. All the movies including the Marvel franchise are now looking at female super heroes. The more conversations we have on this topic, no doubt all the future will be gearing towards gender neutrality. Hence the conclusion of the article steers away from making the article a mouth piece for feminism as AI and every other field are now taking active interest in bringing gender neutrality with an active interest!
Thank you for writing in and it’s great to interact with readers! Pl do keep reading and feeding back as it is absolutely vital for enthusiasts like me to delve deeper in the technology space!
Uma
The article is awesome! Overall, very well researched and written.
Thank you for reading and writing in Lakshmi!
Uma
Great work, Uma. Love the insights!
Good one
Thank you Sarita T
It raises some valid points. Part of the motivation behind the work my organization does is also to ensure that there are more women involved across the board in the health sector, so that the decisions and outputs are not suitable for the ‘reference man’ alone! https://www.womenlifthealth.org/
Interesting! Thank you for your insight!