Answer:
Very good question! There will be human supervision embedded in such workflows, in addition to robots having a 'Perception-Action-Decision Making' system that can identify people and medications with a high degree of accuracy to minimize risks. In such critical sectors, the robotics and automation needs to have Accountability (explainable AI), Responsibility (a chain of actors- Humans, Robots) and Transparency (in Data, Privacy, Decision making); in short called the ART principle. But also remember, humans are not 'error-free', and in some repetitive, laborious tasks robots might make less errors (Robots don't get fatigued, stressed, distracted or need to have coffee/toilet breaks... and can work 24/7 including tough nightshifts without sleeping!).

Answer:
Picking strawberries, tomatoes, Bell Peppers, Salad Crops, Mushrooms! We are developing robots that can assemble sandwiches and fruit salads in food production lines, robots that can inspect nuclear facilities as a part of the National Centre for Nuclear Robotics, Robots that can assist in Carehomes, Microbots for medical applications. We are doing a lot of interdisciplinary research in robotics collaborating with psychologists (human-robot collaboration, infant cognition), neuroscientists (brain inspired models, neural disorders), farmers, medical professionals, fashion designers (wearable technologies) just to mention a few!

Answer:
Several...

• Humanlike learning (in contrast to machine learning)
• Dexterity in the hands to 'act'
• Social intelligence to 'interact'
• From 'Sense' to 'Commonsense' and the ability to deal with new, unknown situations
• Explainability (beyond black boxes)
• Customised and user-centric assistance (needs of every user is different)
• Innovative Robot user interfaces/communication that can accelerate adoption of technology.

At the same time, we could modify our own environments (farms, hospitals, factories, homes) to be more amenable to robotic assistance/automation (at Essex we are working a lot in this direction).

Answer:
Besides the vast applications of robotics, robots are also very valuable research tools to understand ourselves and learn how the brain works (by building robotic models). This approach is called reverse engineering (and developmental robotics). In simple terms, consider that when we try to make a robot see, act, learn, decide, interact or communicate we engineers are solving the same problem that our brain solves to make us see, act, learn, interact and communicate! In this sense, understanding intelligence and building intelligent robots are two sides of the same coin.

Answer:
They could be used to test out revision cards with and go over topics that you need explaining again by showing you how to access relevant YouTube videos. They could even help keep you motivated when revising by taking away any distractions (phones, games controllers etc), bringing you snacks/drinks whilst you're working and also let you know when you need a break for some fresh air!

Answer:
Definitely! We are already seeing early use cases of this with companionship robots for people who are perhaps living alone and want some company. Robots in care homes could be used to help deliver food to people's rooms, help organise people's medicines and maybe even entertain residents by reading books or playing chess with them.

Answer:
Well, strawberry picking for one! Also, jobs that are dangerous or repetitive, like maintaining and decommissioning nuclear reactors, sorting recycling or Marine surveillance.

Answer:
For the likes of robotic surgery, they can be very precise, meaning the size of the incision during an operation can be smaller making recovery time shorter. Robots can also be remotely operated from anywhere in the world, allowing specialists to reach a wider population.

Answer:
Yes, in principle this is possible. Though it is strictly speaking true that a robot/machine only follows the instructions we've given it, those instructions may include instructions on how to learn new things from observation/sensing its environment. There's no way to predict what new things the machine will learn in this way; the new knowledge it acquires is beyond our control. Note also that because modern computers store their instructions as data in memory, they can amend their own instructions. All this suggests that at least in principle there's nothing to rule out a robot with a mind of its own. Many of the arguments around these questions were already made by Alan Turing in his 1950 paper "Computing Machinery and Intelligence".

Answer:
Our ArguBots have access to a very large database of arguments that were crowdsourced (provided by lots of people). When the ArguBot is discussing a particular topic, it can compare your contributions with those in the database and find the ones that are closest to it. Since it knows what side the contributions in the database are on (when the database was crowdsourced, the contributors were asked to state which side of the argument they were contributing to), the ArguBot can then make an educated guess about your contribution.