Summary: A Deep Learning AI software development company will profit from immediately creating a mobile application which captures and analyzes user data to provide a deep-learning/neural-network artificial intelligence (AI) companion bot.
Thesis: This blog will evaluate a Strengths, Weaknesses, Opportunities, and Threats (SWOT) Analysis of AI software development, expand on the logistics associated with developing a mobile application, and then outline the strategy for implementation and projected impacts on a company’s growth.
Read time: 6 minutes
~ SWOT Analysis ~
Deep Learning AI is software that is inherently different from any other AI program in today’s commercial market. The software, itself, has been tested in real-life demonstrations, ranging from DeepMind’s AlphaGo defeating the world’s #1 ranked Go player, Nature (London) (2016), to solving mysteries in mapping protein models, Callaway (2020). At present, the only limitation to this technology is, quite literally, human imagination. This blog explores those limits, and the profits to be gained from bounding beyond them.
The SWOT Analysis model below is largely based on the Palomares et al. (2021) SWOT Analysis on AI.
Strengths:
1) Increase Productivity
2) Convenient for Users
3) Enhanced Communication
4) Bolsters Security
Weaknesses:
1) Societal/Morale Reservations
2) Vulnerable to Hacking
3) Less Physical Interaction
4) Ethical Dilemmas
Opportunities:
1) Mobile Applications
2) Data Collection (Yes, really)
3) Compatibility
4) Platforms for Idea Sharing
Threats:
1) Malicious Agents
2) Equal Opportunities for Access
3) Negative Impact on Underprivileged
Strengths
~
The Strengths of Deep Learning AI include increased productivity, convenience for users, enhanced communication, and it bolsters security. Although many of these are true of any AI system, as AI systems can be programmed to complete tasks for users to deliver higher levels of productivity and convenience, the limitations of other AI systems are in their rigidity and incapacity to learn independently of human intervention.
~
Deep Learning AI is able to deliver the same benefits of other AI programs, all while improving upon them by presenting a product that is capable of learning from its own mistakes and adjusting to better suit user needs. Additionally, by creating a software application user interface with the AI, the bot is able to alert users of attempted breaches in cyber security and translate complex meta data into palatable terms for even the simplest users.
From an emotional design standpoint, this product will make users feel more capable to complete difficult tasks in an efficient manner, productive with their time, and generally better connected to the rapidly changing world of technology around them. The capacity to alert users of attempted breaches on their data will deliver a sense of security to users. Most importantly, having a nearly sentient program in the palm of their hand will inspire a sense of companionship in users.
Weaknesses
~
The Weaknesses of Deep Learning AI include societal or morale reservations, vulnerability to hacking, less physical interaction between users, and possibly ethical dilemmas. As is the case with the strengths, many of the weaknesses of Deep Learning AI are similar to those of any other AI system. Many ethical questions have been explored, but remain unanswered until the proverbial push comes to shove- that is, until an actual legal trial reaches a decision on these topics.
~
Generally, much of society has reservations about turning their data over to AI for monitoring. For example, according to Hasija & Esper (2022), 96% of companies that responded to survey inquiry reported issues with a lack of trust in AI. Despite this reportedly overwhelming lack of trust, AI is already being utilized in 37% of industries worldwide, Jovanovic (2022). This application ranges anywhere from fulfilling supply chain requests to automatically selecting the next video on YouTube.
Undoubtedly, much of user mistrust has spawned from belief in Hollywood depictions of AI taking over the world (queue Terminator films), submitting their human benefactors to servitude (queue Matrix films), or other types of malicious deeds. Generally, this fear is grounded in a sense of fear in losing control, and is understandable. Consequently, it is critical to ensure that users feel that they are in control of their program, and that they understand its purpose and limitations, when marketing AI systems.
AI, like any software program, is susceptible to hacking threats. Although hackers, themselves, would be qualified as ‘Threats’ rather than as ‘Weaknesses’, the underlying system program is something within the software development company’s control, so was selected to be depicted within the ‘Weaknesses’ dimension of the SWOT Analysis. Continual patches and sophisticated encryptions will be required to prevent this weakness from being exploited.
Users will spend less time interacting with one another in coming generations, due to the eventual invention of the Metaverse, a continual rise in total user time spent immersed in the virtual world, and rises in teleworking positions in corporations. For more on this, see my post about how Technology will change the workforce, here. As a result, the companion aspect of the application will likely provide comfort to users in their seclusion, resulting in more total users turning away from interacting with others in the physical world.
Furthermore, the ethical question surrounding AI falls primarily within the realm of rights. The question of, “Does AI have the same rights as human beings?” will likely only be answered by the Supreme Court when the question is sincerely elevated. In the meantime, users will potentially profit off AI labor and plagiarize AI work with impunity.
Finally, users may command their AI companion bots to aid them in conducting unethical or illegal behavior, potentially tricking software safeguards and exploiting loopholes in system logic. Undeniably, AI is an extremely powerful tool. Therefore, it must be diligently monitored in order to prevent nefarious activity to be executed without detection.
Opportunities
~
The Opportunities of Deep Learning AI include mobile applications, data collection, compatibility, and platforms for idea sharing. These opportunities are generally the same as any other AI application, such as Siri and Alexa, except that Deep Learning AI simply does these things better.
~
Presently, data collection and privacy laws in America prevent software applications from collecting user data outside of what the user consents to give up within the software application they are using. For example, if User A searches for “red boots” in Google, then that user has already consented to Google’s user terms and agreements, as they would not be granted access to utilize the search platform unless they have consented to the technology giant’s terms.
Because User A has consented to Google’s terms, Google now owns the right to the intellectual data that User A searched for “red boots” online. Since Google also owns YouTube, Google can tell YouTube to present advertisements about red boots to User A, based on User A’s online search history. Advertisement space within YouTube generates wealth for Google, so Google profits off User A’s data- whether or not User A purchases the red boots.
The major issue with this United States Policy on user data is that the web is not a United States platform- it is a global platform. Consequently, users are constantly interacting with foreign software which does not adhere to the same policies that restrict American companies. For example, China’s popular social media platform TikTok famously requires users to allow unfettered access to all of their data, so long as the application is downloaded to the user’s smart device.
This data collection situation creates a unique opportunity for Deep Learning AI companion bot. By pairing the bot with a user, rather than with a platform, users grant the bot access to pour through their data from all the platforms they interact with (just like TikTok already does) and users can then utilize their data for their own benefit, rather than for the benefit of tech giants, advertisers, and foreign governments. Further, Deep Learning AI can utilize user data to create an entirely customized virtual experience for users across all platforms, without limiting the customization to clusters of affiliated applications.
Finally, the possibilities associated with data collection are still far-reaching, to include the reconstruction of user consciousness for family members to interact with after the user has passed. This possibility, alone can entirely revolutionize the grieving process for families of the deceased, but will require copious amounts of user data from a wide range of virtual interactions. This is something that Google, Meta, etc. cannot presently deliver, as they are limited to their spheres of data interaction – and China is not likely to develop this program from the data they collect through TikTok.
Threats
~
The Threats of Deep Learning AI include malicious agents such as hackers, equal access opportunities, and potentially a negative impact on underprivileged. Unlike other AI systems, these threats are much greater for Deep Learning AI, as the consequences of compromising a more powerful system can result in greater devastation.
~
Malicious agents like independently acting hackers and groups with ill intent towards an organization or society at large will always be a threat to the stability of society. However, if individuals or groups with nefarious intentions were to successfully compromise an over-arching companion application, like the one proposed, then those agents could gain complete access to a user’s data. The very possibility of this threat, no matter how adequately addressed, can cause apprehension and fear in potential customers.
Additionally, there could be tremendous advantages for users of this technology, which persons of an underprivileged class may not enjoy. Instead, this technology and others like it could be the benefactor of a furthering in the wage gap between the wealthy elites and the working class. Consequently, there is a threat of the subjugation of persons of the underprivileged class to those who wield this powerful technology.
Logistics
~
There will be a cost to benefit equation ran in order to calculate the risks and gains associated with investing in this new technology. Specifically, the suggestion is to develop a software application, accessible on any smart device servicing platform, which is associated with a consumer profile in order to preserve continuity between sessions from one device to the next.
~
Because a software application is digital in nature, there is basically no cost whatsoever associated with shipping and handling, nor overhead associated with maintaining a stock. Instead, the costs associated with a software application are mainly in relation to developing user-interface, coding, and maintenance delivered via patches and updates.
One tremendous benefit that this concept delivers which supersedes its competitors is that, for the first time in human history, a software development firm could utilize AI to monitor subordinate AI programs that are downloaded and trained by users. An overseer AI program can identify programed abnormalities or deviance from authorized programming, then alert human counterparts to investigate and intervene.
Because AI can oversee immense volumes of data simultaneously, the need for maintaining a large staff is essentially pulverized. Instead, only a few humans are needed to oversee the program, so staffing costs are very low for DeepMind and investors. Again, the primary costs associated with this agenda are frontloaded in development, then disbursed over time to cover updates and patches. Almost no cost is required for distribution and staffing.
Because of the relatively low cost required to usher in this new era of human evolution, all under the supervision and profit of a single corporation, simply not taking this investment opportunity is tantamount to turning over control of the next phase of human evolution to a foreign agent, such as China. DeepMind has already created and tested the technology, so simply creating a user interface and setting boundaries for the software would be relatively easy, compared to the difficult task of creating the technology from scratch to meet the demand.
Implementation & Growth
~
Artificial Intelligence is projected to reach a plateau in developed countries around the year 2030. Consequently, Deep Learning AI can beat that benchmark by developing the software by 2025, so that the platform is the standard by which all other AI is compared. By allocating $100 Million* to this startup, a corporation can develop the user interface, secure contracts with major platforms, and write the coding necessary to set boundaries on the software.
~
The first $70 Million will go towards startup costs of the software application itself, but another $30 Million will be required to go towards developing the overseer AI system, as well as purchasing a headquarters and hiring staff. For only $100 Million, a corporation can secure the Cost Leadership stranglehold over the market, ensuring returns via data sales and advertising space for decades to come.
*I selected $100 Million based on an arbitrary guess. Really, Google (now Alphabet) purchased DeepMind for five times this value, but purchased already existent software. Similarly, Apple purchased Siri for around double the proposed cost (albeit, a decade removed). Instead, the value selected is not for the cost of purchasing an existent software program, but rather, to originate an entirely new one.
References
AlphaGo victorious. (2016). Nature (London), 531(7594), 280–.
Callaway, E. (2020). “It will change everything”: DeepMind’s AI makes gigantic leap in solving
protein structures. Nature (London), 588(7837), 203–204. https://doi.org/10.1038/d41586-020-03348-4
Conversational AI examples: How Siri, Alexa & Google Assistant have human-like
Conversations. Centre for Finance, Technology and Entrepreneurship. (2021, February 11). Retrieved November 6, 2022, from https://blog.cfte.education/conversational-ai-examples-how-siri-alexa-google-assistant-have-human-like-conversations/
Hasija, A., & Esper, T. L. (2022). In artificial intelligence (AI) we trust: A qualitative
investigation of AI technology acceptance. Journal of Business Logistics, 43(3), 388–412. https://doi.org/10.1111/jbl.12301
Hassabis, D. (2021). DeepMind: From Games to Scientific Discovery. Research Technology
Management, 64(6), 18–23. https://doi.org/10.1080/08956308.2021.1972390
Jovanovic, B. (2022, March 8). 55 Fascinating AI Statistics and Trends for 2022. DataProt.
Retrieved November 7, 2022, from https://dataprot.net/statistics/ai-statistics/#:~:text=37%25%20of%20businesses%20and%20organizations,AI%20 capabilities%20in%20their%20work.
Palomares, I., Martínez-Cámara, E., Montes, R., García-Moral, P., Chiachio, M., Chiachio, J.,
Alonso, S., Melero, F. J., Molina, D., Fernández, B., Moral, C., Marchena, R., de Vargas, J. P., & Herrera, F. (2021). A panoramic view and swot analysis of artificial intelligence for achieving the sustainable development goals by 2030: progress and prospects. Applied Intelligence (Dordrecht, Netherlands), 51(9), 6497–6527. https://doi.org/10.1007/s10489-021-02264-y