Ada Lovelace

Augusta Ada King, Countess of Lovelace (born Augusta Ada Byron; 10 December 1815 to 27 November 1852), usually called Ada Lovelace, was an English mathematician known for her work on Charles Babbage's proposed Analytical Engine. In 1843 she published an English translation of a French memoir describing the machine, together with a set of seven explanatory notes (labeled A through G) that ran to about three times the length of the original text ^[1][2]. Those notes are the reason her name recurs in the history of computing and of artificial intelligence.

Two parts of the notes are cited most often. Note G lays out a step-by-step method by which the engine could compute the Bernoulli numbers, a worked example that is frequently described as the first algorithm published for a general-purpose computing machine, though the exact credit due to Lovelace as against Babbage has been debated for more than a century ^[1][3]. Note A argues that the engine could operate on any objects whose relations could be expressed symbolically, not only on numbers, and used the example of composing music ^[1]. Note G also contains the sentence that Alan Turing would later name the "Lovelace objection": that the machine "has no pretensions whatever to originate anything" ^[1][4].

Lovelace's ideas re-entered the modern debate about machine intelligence through Turing's 1950 paper, which answered her objection directly ^[4]. Her name was later attached to a formal test of machine creativity proposed in 2001, to a British research institute on data and AI founded in 2018, and to a graphics-processor architecture that NVIDIA released in 2022 and used across many of its AI accelerators ^[5][6][7]. This article covers her life and mathematical education, the 1843 translation and notes, the historiographic dispute over the "first programmer" label, and the later uses of her name in AI.

Life and mathematical education

Ada Byron was the only legitimate child of the poet George Gordon Byron (Lord Byron) and Anne Isabella Milbanke ^[8][3]. She was born in London on 10 December 1815. Byron left England in April 1816, when Ada was a few months old, and died in Greece in 1824 without seeing her again ^[8]. Her mother, herself educated in mathematics, arranged for Ada to be tutored in the subject from an early age, in part out of a stated wish to steer her away from what the family regarded as her father's temperament ^[6][8].

Her tutors and correspondents included the social reformer and mathematician William Frend, the physician and mathematics tutor William King, and the science writer Mary Somerville, whom she met in 1833 or 1834 and who kept up a scientific correspondence with her for years ^[8][3]. Somerville, the translator of Laplace's celestial mechanics into English, was one of the connections that brought Ada into scientific circles ^[3]. From around 1840 to 1841 Ada studied more advanced mathematics by letter with Augustus De Morgan, the first professor of mathematics at London University and a distinguished logician ^[3][8].

The depth of that training became the central evidence in later assessments of her ability. In a confidential letter to Lady Byron in January 1844, De Morgan wrote that Ada's "power of thinking . . . from the beginning of my correspondence with her, has been something so utterly out of the common way for any beginner, man or woman," and he rated her facility above that of Mary Somerville ^[3]. A detailed reappraisal of the Lovelace-De Morgan letters, published by Christopher Hollings, Ursula Martin, and Adrian Rice in 2017, reordered the surviving archive material and argued against earlier claims that Lovelace was mathematically weak, presenting instead a more nuanced assessment of her abilities ^[9].

In 1835 Ada married William King, a baron who was created Earl of Lovelace in 1838; she thereby became Countess of Lovelace, which is why she is known as Ada Lovelace rather than by her longer formal style ^[3][8]. The couple had three children. She died in London on 27 November 1852 of cancer, aged 36, and was buried beside her father at the Church of St Mary Magdalene in Hucknall, Nottinghamshire ^[8][10].

Babbage and the Analytical Engine

Ada first met Charles Babbage at a party in 1833, when she was seventeen, and shortly afterward saw the working portion of his Difference Engine, a special-purpose calculator that tabulated polynomial functions by the method of differences ^[11][3]. The Difference Engine repeated a single operation, addition, across a set of columns, and could evaluate polynomials up to the sixth degree; it impressed her, and she grasped its significance more readily than most visitors did ^[11][1].

Babbage soon set that project aside for a more ambitious design, the Analytical Engine, a general-purpose machine that he continued to develop for the rest of his life but never built ^[10]. Its plan separated a "store" (memory, holding numbers on columns of geared wheels) from a "mill" (the arithmetic unit), and it was to be controlled by punched cards on the principle Joseph-Marie Jacquard had used to drive patterned looms ^[1]. The design included the ability to repeat groups of cards and to branch depending on a computed result, features that correspond to loops and conditionals in later programming ^[1][10]. Because it combined programmable control with a general set of operations, the Analytical Engine is generally regarded as the first design for a machine capable of general-purpose computation ^[10].

The engine's first published description came not from Babbage but from Luigi Federico Menabrea, an Italian military engineer who later became prime minister of Italy. After hearing Babbage lecture in Turin in 1840, Menabrea wrote a short account in French, "Notions sur la machine analytique de M. Charles Babbage," which appeared in the Bibliothèque Universelle de Genève in October 1842 ^[1][12]. It was this memoir that Lovelace translated.

The 1843 translation and notes

Over the winter of 1842 to 1843 Lovelace translated Menabrea's memoir, roughly 8,000 words, and showed the result to Babbage in the spring of 1843 ^[3]. Accounts differ slightly on how the project grew. Babbage, in his own memoir, recalled asking why she had not written an original paper on a subject she understood so well, and then suggesting that she add notes to the translation, "an idea which was immediately adopted" ^[12]. Other accounts credit the scientist and telegraph inventor Charles Wheatstone, who knew both parties and was collecting contributions for the journal, with prompting the translation in the first place ^[3].

The translation and its seven notes were published in 1843 in Richard Taylor's Scientific Memoirs, volume 3, pages 666 to 731, under the title "Sketch of the Analytical Engine invented by Charles Babbage, Esq." ^[1]. Lovelace signed the notes only with her initials, "A.A.L." ^[1]. The notes together are about three times as long as Menabrea's original and treat, in Babbage's words, "almost all the very difficult and abstract questions connected with the subject" ^[12].

In Note A, Lovelace drew the distinction that later readers found most striking: the Difference Engine could tabulate one class of function, but the Analytical Engine was "the material expression of any indefinite function of any degree of generality and complexity," configured by its Jacquard-style cards ^[1]. She summarized the contrast in a much-quoted sentence: "We may say most aptly, that the Analytical Engine weaves algebraical patterns just as the Jacquard-loom weaves flowers and leaves" ^[1].

Note G and the Bernoulli numbers

Note G, the last and longest note, works through how the engine could compute the Bernoulli numbers, a sequence of rational numbers that arises in the summation of powers of integers and elsewhere in analysis ^[3]. Lovelace chose the example deliberately, as she explained in a letter to Babbage: she wanted to show "how an implicit function may be worked out by the engine, without having been worked out by human head & hands first" ^[3]. The note includes a large fold-out table titled "Diagram for the computation by the Engine of the Numbers of Bernoulli," which traces, operation by operation, how the engine would compute the value she labeled B7 ^[1][3].

The plan is organized as a sequence of numbered operations acting on numbered variable columns. Lovelace noted that for the general case the computation used twenty-five operation cards, that "these same twenty-five cards suffice for the successive computation of all the Numbers" of Bernoulli, and that a group of operations (which she numbered 13 through 23) is repeated as the calculation moves from one Bernoulli number to the next ^[1]. That repeated group is the reason the example is often described as containing an early loop, although Lovelace had no symbolic notation for loops and indicated the repetition in prose and with braces; indexed loop constructs of the modern kind appeared only about a century later ^[3].

Babbage supplied the underlying mathematics for this note. In his memoir he wrote that the algebraic working of the different problems was Lovelace's "except, indeed, that relating to the numbers of Bernoulli, which I had offered to do to save Lady Lovelace the trouble," and he added that she had returned it to him "for an amendment, having detected a grave mistake which I had made in the process" ^[12]. The surviving correspondence shows Lovelace working through the details herself: "I am in much dismay at having got into so amazing a quagmire & botheration with these Numbers," she wrote to Babbage during the work ^[10]. Later commentators examining the printed table found a single misplaced minus sign which, once corrected, leaves an algorithm that computes the Bernoulli numbers correctly; one historian reports that the table has been translated into a 65-line FORTRAN program that reproduces the sequence ^[3].

The "first programmer" debate

Whether Note G makes Lovelace the first computer programmer has been argued in both directions, and the scholarship on the point is unusually sharp.

The critical case rests on the fact that the Analytical Engine's operation was already illustrated with tables before her notes. Menabrea's memoir itself contained tables of operations, for example for solving two simultaneous linear equations, and Babbage had prepared example computations for the engine years earlier ^[1][3]. The Babbage historian Allan Bromley concluded that "all but one of the programs cited in her notes had been prepared by Babbage from three to seven years earlier," the exception being the Bernoulli computation, which "was prepared by Babbage for her, although she did detect a 'bug' in it"; Bromley added that "there is no evidence that Ada Lovelace ever prepared a program for the Analytical Engine" ^[3]. Bruce Collier's 1970 Harvard thesis reached a similar verdict, stating that although Lovelace understood Babbage's ideas and "expressed them well in her notes," it is "equally clear that the ideas were indeed Babbage's and not hers," and that "there is no evidence that she advanced the design or theory of it in any way" ^[13]. Dorothy Stein's 1985 biography called her "a figure whose achievement turns out not to deserve the recognition accorded it," and the survey Computer: A History of the Information Machine states that "most of the technical content and all of the programs in the Sketch were Babbage's work" ^[3].

More recent work pushes back without simply reversing the charge. The historian Thomas Misa argues that the question "is not a zero-sum game, where any credit added to Lovelace somehow detracts from Babbage," and that both contributed importantly to the Sketch and its notes ^[3]. On this reading, the underlying formula for the Bernoulli numbers may well have been Babbage's, but "the transformation of the general formula into a step-by-step algorithm remains Ada's achievement, as the letters clearly indicate" ^[3]. Even the definitional problem cuts against the tidy label: because the concept of writing a program for a stored-program computer dates from decades later, describing Lovelace (or anyone in 1843) as the "first programmer" is anachronistic, a point on which both critics and defenders now largely agree ^[14]. Misa's own conclusion is that, whatever one calls it, "Lovelace created a step-by-step elemental sequence of instructions, that is, an algorithm, for computing the series of Bernoulli numbers" that was intended for the Analytical Engine ^[3].

Beyond number: the symbol-manipulation insight

The idea in the notes that reaches furthest toward modern computing is not the Bernoulli program but a claim in Note A about what a computing machine is for. Lovelace defined an "operation" broadly as "any process which alters the mutual relation of two or more things," and described the engine as an embodiment of a "science of operations" that is distinct from the numbers it happens to work on ^[1]. From that starting point she argued that the machine "might act upon other things besides number, were objects found whose mutual fundamental relations could be expressed by those of the abstract science of operations." Her illustration was music: if the relations of pitched sounds "were susceptible of such expression and adaptations, the engine might compose elaborate and scientific pieces of music of any degree of complexity or extent" ^[1].

That passage is frequently read as an early statement that a computer manipulates symbols in general, of which arithmetic is only one case, an idea central to later work on symbolic AI and, more broadly, to any machine that processes text, images, or sound. Stephen Wolfram, reviewing the primary documents for Lovelace's bicentennial, argued that in trying to describe Babbage's machine abstractly she "ended up exploring and articulating something quite recognizable as the modern notion of universal computation," and called her "the first person ever to glimpse with any clarity" that idea ^[10]. Wolfram is careful to note that the physical engine was entirely Babbage's design, and that Babbage did not think in terms of universal computation himself ^[10].

The Lovelace objection and Turing's reply

Note G contains the passage that connects Lovelace most directly to the philosophy of artificial intelligence. Guarding against "exaggerated ideas" about the engine, she wrote:

"The Analytical Engine has no pretensions whatever to originate anything. It can do whatever we know how to order it to perform. It can follow analysis; but it has no power of anticipating any analytical relations or truths. Its province is to assist us in making available what we are already acquainted with." ^[1]

Alan Turing took up this statement in his 1950 paper "Computing Machinery and Intelligence," published in the journal Mind, volume 59, pages 433 to 460 ^[4]. In the section he titled "Lady Lovelace's Objection," Turing quoted her line that the engine "has no pretensions to originate anything" and noted the surrounding words, then set out to answer it as part of his broader defense of the question that became the Turing test ^[4]. He first cited the physicist Douglas Hartree, who had observed in 1949 that Lovelace's statement did not prove that no machine could ever "think for itself," only that the machines built or projected at the time gave no evidence of doing so ^[4]. Turing agreed, and added that the Analytical Engine was in fact a universal digital computer, so that with adequate storage and speed it could by suitable programming be made to imitate any other discrete-state machine; the evidence available to Lovelace, he wrote, simply "did not encourage her to believe" that machines had such a property ^[4].

Turing then reframed a stronger form of the objection, "that a machine can never take us by surprise," and answered it flatly: "Machines take me by surprise with great frequency," he wrote, attributing his critics' contrary view to an unexamined assumption that all the consequences of a fact spring into the mind at the same time as the fact itself ^[4]. The exchange, one figure writing in 1843 and another answering in 1950, is a standard reference point in discussions of machine creativity and originality and of whether systems such as modern generative AI tools can be said to produce anything genuinely new.

The Lovelace Test and its successors

Lovelace's objection also gave its name to a family of proposed tests of machine intelligence framed around creativity rather than conversation. In 2001 Selmer Bringsjord, Paul Bello, and David Ferrucci proposed the "Lovelace Test" in the journal Minds and Machines ^[15]. In their formulation an artificial agent a, designed by a human h, passes the test only if a produces an output o; a's production of o is the result of processes the agent can repeat, rather than a fluke hardware error; and h, or anyone who knows what h knows and has h's resources, cannot explain how a produced o ^[16]. The test was meant to be demanding by design, because a human who built the system is normally able to account for its outputs.

In 2014 Mark Riedl of Georgia Tech proposed the "Lovelace 2.0 Test," arguing that the original was effectively unbeatable and hard to apply ^[16]. His version asks an artificial agent a to create an artifact o of a type t (for example a story, poem, or picture) that satisfies a set of constraints C expressible in natural language; a human evaluator h, having chosen t and C, judges whether o is a valid instance of the type and meets the constraints, while a human referee r checks that the combination of t and C is not unrealistic for an average person ^[16]. Because the evaluator can keep imposing harder constraints, the design is meant to let judges compare systems and probe their limits rather than deliver a single pass-or-fail verdict ^[16].

The framing has continued into the era of large generative models. A 2025 study by Ewelina Gajewska, posted to arXiv as "The Lovelace Test of Intelligence: Can Humans Recognise and Esteem AI-Generated Art?", ran a Lovelace-style experiment in which people with backgrounds in cognitive and computer science tried to tell AI-generated paintings from human-made ones and to rate their aesthetic value. The study reported that participants could not reliably distinguish the two, performing no better than chance under some conditions, and that the AI-generated works were rated as aesthetically comparable to human-made ones ^[17].

Namesakes in modern AI

Lovelace's name is attached to several contemporary AI-related projects. The most widely used is NVIDIA's Ada Lovelace GPU architecture, announced on 20 September 2022, with the first consumer card, the GeForce RTX 4090, becoming available on 12 October 2022 ^[7]. NVIDIA's architecture whitepaper states plainly that the design is "named after mathematician Ada Lovelace, who is often regarded as the world's first computer programmer" ^[20]. The flagship AD102 processor is built on TSMC's 4N process and contains 76.3 billion transistors, 18,432 CUDA cores, 144 third-generation ray-tracing cores, and 576 fourth-generation tensor cores, the last of which perform the matrix arithmetic used in deep learning ^[20]. Ada Lovelace parts extend well beyond gaming into AI infrastructure: the data-center L40 and L40S accelerators use the same architecture, the L40S offering 48 GB of memory and 568 Tensor Cores, and are marketed for training and inference as well as graphics ^[18]. NVIDIA's GPU generations follow the practice of naming architectures after scientists; the Ada Lovelace line was succeeded by Blackwell, announced with the GeForce RTX 50 series on 6 January 2025 ^[19].

Her name also marks work on AI governance. The Ada Lovelace Institute, an independent research body based in the United Kingdom, was established by the Nuffield Foundation in early 2018 in collaboration with the Alan Turing Institute, the Royal Society, the British Academy, the Royal Statistical Society, the Wellcome Trust, Luminate, techUK, and the Nuffield Council on Bioethics ^[5]. Its stated mission is to ensure that data and AI "work for people and society," and it describes itself as connected to Lovelace's legacy of intellectual rigor ^[5]. Together with the NVIDIA Ada Lovelace architecture, these uses keep her name current in a field whose central questions, about what a machine can be ordered to do and whether it can originate anything, she helped to pose.

See also

• Alan Turing
• Turing test
• History of artificial intelligence
• NVIDIA Ada Lovelace
• Artificial general intelligence

References

1. ^Menabrea, L. F. (trans. Ada Lovelace), "Sketch of the Analytical Engine invented by Charles Babbage, Esq., with notes by the translator," Scientific Memoirs, vol. 3 (1843), pp. 666-731. Reproduced by Classics in the History of Psychology, York University. psychclassics.yorku.ca/...lovelace
2. ^Menabrea, L. F. (trans. Ada Lovelace), "Sketch of the Analytical Engine" (main memoir text). Classics in the History of Psychology, York University. psychclassics.yorku.ca/...menabrea
3. ^Misa, Thomas J., "Charles Babbage, Ada Lovelace, and the Bernoulli Numbers," in R. Hammerman and A. L. Russell (eds.), Ada's Legacy (ACM Books, 2015). Preprint at arXiv:2301.02919. arxiv.org/...2301.02919
4. ^Turing, A. M., "Computing Machinery and Intelligence," Mind, vol. 59, no. 236 (October 1950), pp. 433-460. doi.org/...LIX.236.433 (full text: courses.cs.umbc.edu/...turing.pdf )
5. ^Ada Lovelace Institute, "About." adalovelaceinstitute.org/about
6. ^Ada Lovelace Institute, "About Ada Lovelace" (biographical summary on the About page). adalovelaceinstitute.org/about
7. ^NVIDIA, "GeForce RTX 40 Series Graphics Cards," announcement of 20 September 2022 (RTX 4090 available 12 October 2022). nvidia.com/...-series-graphics-cards-announcements
8. ^O'Connor, J. J., and Robertson, E. F., "Augusta Ada King, countess of Lovelace," MacTutor History of Mathematics Archive, University of St Andrews. mathshistory.st-andrews.ac.uk/...Lovelace
9. ^Hollings, C., Martin, U., and Rice, A., "The Lovelace-De Morgan mathematical correspondence: A critical re-appraisal," Historia Mathematica, vol. 44 (2017), pp. 202-231. doi.org/...j.hm.2017.04.001
10. ^Wolfram, Stephen, "Untangling the Tale of Ada Lovelace," Stephen Wolfram Writings, 10 December 2015. writings.stephenwolfram.com/...ale-of-ada-lovelace
11. ^Computer History Museum, "Ada Lovelace." computerhistory.org/...adalovelace
12. ^Babbage, Charles, Passages from the Life of a Philosopher (London: Longman, Green, 1864). Project Gutenberg edition. gutenberg.org/...pg57532.txt
13. ^Collier, Bruce, "The Little Engines That Could've: The Calculating Machines of Charles Babbage," Ph.D. thesis, Harvard University (1970). robroy.dyndns.info/collier
14. ^Misa, Thomas J., "Comment on: Was Ada Lovelace Actually the First Programmer?" (2022). arXiv:2402.00749. arxiv.org/...2402.00749
15. ^Bringsjord, S., Bello, P., and Ferrucci, D., "Creativity, the Turing Test, and the (Better) Lovelace Test," Minds and Machines, vol. 11, no. 1 (2001), pp. 3-27. doi.org/...A:1011206622741
16. ^Riedl, Mark O., "The Lovelace 2.0 Test of Artificial Creativity and Intelligence" (2014). arXiv:1410.6142. arxiv.org/...1410.6142
17. ^Gajewska, Ewelina, "The Lovelace Test of Intelligence: Can Humans Recognise and Esteem AI-Generated Art?" (2025). arXiv:2509.11371. arxiv.org/...2509.11371
18. ^NVIDIA, "L40S GPU for AI and Graphics Performance" (product page and specifications). nvidia.com/...l40s
19. ^NVIDIA, "NVIDIA Blackwell GeForce RTX 50 Series Opens New World of AI Computer Graphics," 6 January 2025. nvidianews.nvidia.com/...d-of-ai-computer-graphics
20. ^NVIDIA, "NVIDIA Ada GPU Architecture" whitepaper (V2.02). images.nvidia.com/...nvidia-ada-gpu-architecture.pdf